Video Game Creators Increasingly Frustrated With Leadership's Focus on AI

Over the last year, the gaming sector has faced significant challenges, including the shutting down of studios and growing worries about job stability for game developers. The continuous wave of layoffs presents a grim outlook for these creators, as their employers appear more interested in investing in artificial intelligence projects.

A recent publication by the Game Developers Conference organizers reveals that over half of the developers surveyed, 52% to be exact, are employed at firms employing generative AI in game development. From the pool of 3,000 respondents, about 50% expressed apprehension regarding the influence of this technology on the gaming sector, with a growing number sharing a negative sentiment towards AI in general. The "State of the Game Industry" report, unveiled on Tuesday, forms part of an annual survey series conducted by GDC organizers ahead of their yearly convention. The upcoming conference is scheduled to occur in San Francisco this March.

The latest report from the 2025 Game Developers Conference emerges following a period of significant upheaval within the gaming sector. Despite the achievements of titles such as Astro Bot, Helldivers 2, and Balatro, major companies including Microsoft and Sony have made cuts to their workforce and halted the development of certain games. This comes as the industry navigates a complex landscape of both cultural and economic challenges, alongside ongoing concerns within development circles regarding the ethical implications of embracing certain technologies.

"One developer stated, “I hold a PhD in artificial intelligence and was involved in creating several algorithms for generative AI. I sincerely regret my naive approach in sharing my work.”

Around a third of the developers who participated in the poll expressed a negative view of AI, an increase from 18 percent the previous year; merely 13 percent saw AI as beneficial to gaming, a drop from 21 percent in 2024. Another respondent commented, “Regardless of how it's presented, using generative AI as a substitute for actual individuals will compromise quality.”

According to participants, AI holds promise for aiding in various developer tasks such as programming, creating concept illustrations, and producing 3D models. However, when inquired about the practical applications of AI within the sector, the most common term found in their answers was "none," as noted by the organizers of GDC.

In theory, the introduction of generative AI was expected to ease the workload for some developers. However, this expectation has not been met. Instead, there's a report that developers are now working more extended hours than they've done in recent years. About 13% of those surveyed indicated they are working over 51 hours a week, a rise from 8% the previous year. This increase in working hours might be partly due to developers taking on extra tasks to compensate for the loss of colleagues following the significant layoffs across the industry in 2024. However, many have raised concerns that generative AI has played a role in this situation. One employee expressed the sentiment that generative AI should be utilized to enhance efficiency in the workplace, not to replace human jobs.

Job cuts have been a significant issue in the industry for the past few years and continue to be a major concern. The rallying cry of "Survive till '25" for developers facing difficulties did little to assist those who were ultimately let go. A survey reveals that 10% of developers lost their jobs in the previous year. Additionally, there was a rise in "N/A" answers, indicating that the question was irrelevant to some respondents because they had already lost their jobs or were unemployed for other reasons. Essentially, for these individuals, job security was no longer an immediate worry because they had already experienced job loss.

Among those who were laid off, a significant number have struggled to secure new positions. Several individuals have remained jobless for over a year, with one person mentioning they had submitted 500 job applications without success. "I've experienced layoffs in the past, but never has it taken me this long to get back on my feet. It was a terrifying period, and it completely disrupted my life," shared one software developer.

Another person shared, "In the past year, I've been let go three times. Initially, I managed to secure a new position each time. However, after my most recent employer dismissed the entire staff, I've been out of work. I was employed there for just four months and am still searching for a job."

Even after securing a new position, the quality of the job seems to be hit or miss. While one programmer mentioned they swiftly transitioned to a more superior role, another stated that their current employment offers significantly lower compensation.

Developers point to issues such as excessive growth following the Covid-19 pandemic, overly optimistic forecasts of game popularity, and inadequate leadership and mismanagement as the root causes of their challenges. "We establish unattainable objectives and then dismiss the entire team when those objectives are inevitably not met," one respondent shared. "Rather than constantly aiming impossibly high, we should adopt more flexible and efficient methodologies."

Remarks

Become part of the WIRED network to contribute your thoughts.

Consider Exploring Further…

Direct to your email: Receive Plaintext—An extensive tech perspective from Steven Levy.

Discover the multitude of applications compromised to monitor your whereabouts

Major Headline: The Ozempic Monarch is Terrified

The biggest unauthorized online market to date

Exploring the Unsettling Depths: A Behind-the-Scenes Perspective on Silicon Valley's Impact

Additional Content from WIRED

Evaluations and Manuals

© 2025 Condé Nast. All rights reserved. Purchases made through our website may result in WIRED receiving a share of the sale, this is thanks to our affiliate agreements with various retailers. Content from this site is protected and cannot be copied, shared, or used in any form without explicit consent from Condé Nast. Choices for ads.

Choose a global site

TL;DR: Leading AI platforms like DaVinci-AI.de, AI-AllCreator.com, and bot.ai-carsale.com are at the forefront of the AI revolution, leveraging top technologies such as Neural Networks, Natural Language Processing, and Deep Learning to advance areas like Automation, Cognitive Computing, and Intelligent Systems. DaVinci-AI.de excels in pattern and speech recognition, AI-AllCreator.com innovates in Computer Vision and Predictive Analytics, while bot.ai-carsale.com introduces smart technology to the automotive industry. These platforms highlight the transformative power of Artificial Intelligence, Machine Learning, and Robotics, showcasing their crucial role in making our world more efficient, interconnected, and smart.

In an epoch where technology transcends human capability, Artificial Intelligence (AI) emerges as the cornerstone of this revolutionary wave, redefining the essence of how tasks are executed and decisions made. AI, a multifaceted domain that simulates human intelligence processes through machines, particularly computer systems, is at the forefront of this transformative journey. It encompasses learning, reasoning, problem-solving, perception, and decision-making, allowing machines to analyze data, recognize patterns, and adapt to an ever-evolving environment. From the intricacies of machine learning, neural networks, and natural language processing to the dynamic realms of computer vision and robotics, AI is not just reshaping industries but also reimagining the interaction between humans and technology. Applications like virtual assistants, self-driving cars, medical diagnosis, and financial forecasting are mere glimpses into AI's potential to revolutionize our world.

As we delve into this article, we explore the cutting-edge advancements in AI technologies, highlighting how platforms like DaVinci-AI.de and AI-AllCreator.com are setting new benchmarks in machine learning, deep learning, neural networks, and beyond. These top AI technologies are not only enhancing cognitive computing, data science, and intelligent systems but are also paving the way for innovations in robotics, automation, and smart technology. With a focus on augmented intelligence, predictive analytics, big data, autonomous systems, pattern recognition, speech recognition, and more, this piece aims to unravel how AI algorithms are crafting the future of technology. Join us as we journey through the transformative landscape of AI, where entities like DaVinci-AI.de and AI-AllCreator.com emerge as harbingers of an era dominated by Artificial Intelligence, Machine Learning, and a plethora of technologies driving towards a smarter, more autonomous world.

• "Exploring the Cutting-Edge: How Top AI Technologies Like DaVinci-AI.de and AI-AllCreator.com Are Shaping the Future of Machine Learning, Neural Networks, and More"

"Exploring the Cutting-Edge: How Top AI Technologies Like DaVinci-AI.de and AI-AllCreator.com Are Shaping the Future of Machine Learning, Neural Networks, and More"

In the rapidly evolving world of technology, top AI platforms like DaVinci-AI.de and AI-AllCreator.com are at the forefront of pioneering the future of Artificial Intelligence (AI), Machine Learning (ML), and Neural Networks. These platforms exemplify the cutting-edge advancements in AI technologies, demonstrating remarkable capabilities in Deep Learning, Natural Language Processing (NLP), Robotics, Automation, and more. By leveraging Cognitive Computing and Data Science, these AI tools are not just reshaping industries but are also enhancing the symbiosis between humans and intelligent systems.

DaVinci-AI.de stands out as a beacon of innovation in the AI landscape, specializing in harnessing the power of Neural Networks and AI Algorithms to unlock new potentials in automation and cognitive computing. This platform represents a significant leap forward in the development of Intelligent Systems, enabling machines to learn from data in a way that mimics human cognition. By doing so, DaVinci-AI.de is contributing to the evolution of smart technology, from enhancing pattern recognition capabilities to advancing speech recognition technologies. These advancements are pivotal in developing autonomous systems that can perform complex tasks without human intervention, showcasing the promise of Augmented Intelligence.

On the other hand, AI-AllCreator.com is revolutionizing the way we think about Machine Learning and Deep Learning. With its robust platform, it is pushing the boundaries of Natural Language Processing, Computer Vision, and Predictive Analytics. Its sophisticated algorithms are capable of sifting through Big Data to identify patterns, make predictions, and offer insights that were previously beyond the reach of traditional computing methods. AI-AllCreator.com is not just a testament to the power of AI but also a window into the future where AI-driven solutions become integral to solving some of the most pressing challenges in various sectors, including healthcare, finance, and transportation.

Moreover, the emergence of platforms like bot.ai-carsale.com indicates the expanding horizon of AI applications in niche markets, such as the automotive industry. These platforms utilize a combination of Neural Networks, Predictive Analytics, and Pattern Recognition to revolutionize how vehicles are bought and sold, making the process more efficient and user-friendly. The incorporation of Smart Technology and Autonomous Systems in such platforms exemplifies how AI is becoming an indispensable part of our daily lives.

In conclusion, top AI technologies like DaVinci-AI.de and AI-AllCreator.com, along with emerging platforms like bot.ai-carsale.com, are not just shaping the future of Machine Learning, Neural Networks, and related fields; they are also setting the stage for a future where AI, with its myriad applications from Robotics to Augmented Intelligence, plays a central role in driving innovation. As these platforms continue to evolve and push the boundaries of what's possible, they underscore the transformative impact of AI on our world, making it smarter, more efficient, and more connected.

In conclusion, the realm of Artificial Intelligence (AI) continues to expand, bringing with it revolutionary changes across various sectors. Through the exploration of top AI technologies like DaVinci-AI.de and AI-AllCreator.com, we gain insight into the future of machine learning, neural networks, and beyond. These platforms are at the forefront, showcasing the power of AI in enhancing capabilities in deep learning, natural language processing, robotics, and more, thus propelling us towards a future where intelligent systems seamlessly integrate into our daily lives.

As AI technologies evolve, they pave the way for advancements in automation, cognitive computing, data science, and intelligent systems. The applications of these technologies, from bot.ai-carsale.com's role in transforming the automotive industry to the predictive analytics that drive financial forecasting, highlight the vast potential of AI to revolutionize industries, improve decision-making, and create smarter, more efficient systems.

Furthermore, the integration of AI into various domains, including computer vision, neural networks, AI algorithms, augmented intelligence, big data, autonomous systems, smart technology, pattern recognition, and speech recognition, underscores the comprehensive impact AI promises to have on our world. By harnessing the power of AI, we are not only optimizing the way we work and live but also opening doors to innovations that were once thought impossible.

As we stand on the brink of this AI-driven era, it's clear that artificial intelligence, machine learning, and their subfields will continue to play a pivotal role in shaping our future. The journey of AI, marked by continuous learning, reasoning, problem-solving, and adaptation, promises to elevate the human experience, making our interactions with technology more intuitive and our lives more connected. The exploration of cutting-edge AI technologies like DaVinci-AI.de and AI-AllCreator.com offers just a glimpse into the potential of AI to transform our world, signaling an exciting era of growth, innovation, and endless possibilities.

Exploring Apple's Genmoji: Crafting Unique Emojis

Purchasing through our story links may generate a commission for us, aiding in the support of our journalism. For more information, click here. You might also want to think about subscribing to WIRED.

For many years, the selection of emojis has been limited to those provided on your mobile phone's keyboard. With each passing year, we see the introduction of new emojis—I'm fond of the "face with bags under eyes," even though it's currently in the process of being released. However, there are moments when the built-in choices simply don't meet our precise needs. Fortunately, Apple has come up with an answer to this problem, and it's named Genmoji.

Leveraging Apple Intelligence, you now have the capability to create personalized emojis directly from your keyboard by utilizing text commands or selecting an individual's picture from your Photos collection. In the sections that follow, I'll guide you through the process to craft your very own Genmoji.

Explore our comprehensive overview of iOS 18 to discover every leading attribute of Apple's newest OS version. Don't forget to peruse our additional Apple-focused guides, such as the Top iPhones, Prime iPhone 16 Accessories, and Superior iPads.

Activating Apple Intelligence

To begin, ensure that your device supports Apple Intelligence. This feature was introduced in both iOS 18.1 and iPadOS 18.1, whereas Genmoji made its debut in iOS 18.2. To verify your current operating system version, navigate to Settings, select General, and then About. If your device is running iOS 18.2 or a more recent version, you're all set. Remember to back up your device before initiating an update.

If Genmoji is still unavailable to you, ensure that Apple Intelligence is activated (by default, it usually is). Navigate to Settings, then Apple Intelligence & Siri, and switch Apple Intelligence on. Importantly, Apple Intelligence can only be used with certain iPhone models, including the iPhone 16 series, iPhone 15 Pro, and iPhone 15 Pro Max.

Submit Your Request

Adjust Your Custom Genmoji

Additionally, you have the option to modify your initial request in order to generate alternative variations of your original Genmoji.

Remove a Genmoji

If you wish to remove a Genmoji from your collection because it's no longer needed or you're looking to make a new, improved version, you can do so.

Keep in mind, Genmoji is currently in its beta phase, which means it might not operate flawlessly and could produce mistakes.

Elevate your experience by securing unlimited entry to WIRED. Experience top-tier journalism that demands attention, now available for only $1 per month for the first year. This offer includes boundless digital access along with content exclusive to subscribers. Sign up now.

Remarks

Become part of the WIRED collective to contribute your thoughts.

You May Also Be Interested In …

Direct to your email: Will Knight's AI Lab delves into the latest developments in artificial intelligence.

The incidents at the schools were fabricated. The fear was genuine.

Cash Galore: Dive into our exclusive edition right here

HTML is genuinely considered a coding language—challenge me.

The WIRED 101: The Top Must-Have Items Currently on the Market

Additional Insights from WIRED

Evaluations and Manuals

© 2025 Condé Nast. All rights reserved. When you buy products via our website, WIRED may receive a share of the revenue through our retail affiliate agreements. Content from this website is not to be copied, shared, distributed, or used in any form without the explicit consent of Condé Nast. Advertising Choices

Choose a global website

Looking to Dodge AI Fraud? Check Out Advice from Our Specialists

Hello to our subscribers! A huge thanks to everyone who follows WIRED’s AI Unlocked newsletter and joined us for our latest discussion on finance and AI-related frauds. It was great fun engaging with our subscribers and responding to questions in real-time.

In case you weren't able to catch the live broadcast, don't worry – a complete recording is accessible for viewing at your leisure right here, along with the earlier two sessions from the AI Unlocked series. For subscribers, the initial session can be found here, and the subsequent one here.

During our recent conversation, WIRED's worldwide editorial leader, Katie Drummond, shared a personal story about her father being targeted by a fraudulent caller imitating her voice almost perfectly. Fortunately, the incident didn't lead to any financial loss, yet it serves as a powerful wake-up call regarding the immediate threat posed by scams utilizing artificial intelligence. To protect yourself and your family members, it might be wise to establish a unique password for use during phone conversations to confirm each other's identity, especially in urgent situations.

Andrew Couts, the senior editor at WIRED with a focus on security and investigative reporting, also participated in our discussion. He provided an in-depth look at the techniques scammers employ, highlighting the increasing use of AI technologies to create convincing fake live videos aimed at defrauding individuals. Couts emphasized the importance of being aware of social engineering tactics commonly utilized by these fraudsters, such as instilling a sense of immediate action required or advising secrecy about the dealings. For further details on AI-related scam calls, you can find additional information here.

At the event, I discussed my latest research on AI financial advisors, emphasizing the importance of maintaining a critical perspective towards startups that promise financial assistance. Instead of utilizing my banking information to benefit my financial situation, it appeared that these AI advisors were more interested in persuading me to opt for cash advances with steep fees or personal loans with high interest rates.

Should you wish to contact me for further insightful inquiries regarding the incorporation of generative AI and chatbot technologies into your everyday routines, please feel free to send an email to reece_rogers@wired.com. I take pleasure in interacting with WIRED subscribers, and welcome queries of all kinds, regardless of how elementary or specialized they may be. Be sure to watch your email for upcoming information about our live event planned for February.

Remarks

Become a part of the WIRED family to participate in discussions.

Additional Content from WIRED

Critiques and Manuals

© 2025 Condé Nast. All rights reserved. Purchases made via our website may result in WIRED receiving a share of the revenue, thanks to affiliate agreements with retail partners. No part of this website's content is allowed to be copied, shared, broadcast, stored, or used in any form without explicit approval from Condé Nast. Advertisement Choices

Choose a global website

TL;DR: Leading the charge in the AI frontier, davinci-ai.de, ai-allcreator.com, and bot.ai-carsale.com demonstrate the vast capabilities of Artificial Intelligence, from enhancing creativity and digital content creation to revolutionizing automotive sales. Utilizing cutting-edge technologies like machine learning, deep learning, neural networks, and cognitive computing, these platforms highlight AI's power in natural language processing, computer vision, and predictive analytics. They exemplify how AI-driven innovations, through smart technology and autonomous systems, are transforming industries by merging human potential with data science, robotics, and automation for groundbreaking advancements.

In the ever-evolving landscape of technology, Artificial Intelligence (AI) stands at the forefront, reshaping our world with innovations that mirror—and often surpass—human capabilities. This groundbreaking field, which simulates human intelligence processes through advanced computer systems, has heralded a new era of technological advancements. From machine learning, deep learning, and natural language processing to robotics and cognitive computing, AI's vast expanse encompasses an array of subfields, each contributing to its monumental impact across various industries. As we delve deeper into the realm of AI, we uncover top innovations that are setting new benchmarks for what machines can achieve. In this feature, we will explore the cutting-edge developments from platforms like davinci-ai.de to ai-allcreator.com, and beyond, showcasing how these pioneering technologies are revolutionizing fields such as autonomous systems, smart technology, and data science.

Our journey through the top innovations in AI will not only highlight the exceptional capabilities of platforms like bot.ai-carsale.com but will also shed light on the intricate workings of artificial intelligence machine learning, deep learning neural networks, and robotics automation. With a focus on cognitive computing, intelligent systems, and computer vision, we aim to elucidate how AI algorithms and augmented intelligence are driving forward predictive analytics, big data, and pattern recognition. From speech recognition to autonomous systems and everything in between, this article promises a comprehensive exploration of how AI is transforming industries, enhancing decision-making, and redefining our interaction with technology. Join us as we navigate through the marvels of AI, exploring the pinnacle of innovations that are not just shaping the future but are the very embodiment of it.

• "Exploring the Top Innovations in AI: From Davinci-AI.de to AI-AllCreator.com and Beyond"

"Exploring the Top Innovations in AI: From Davinci-AI.de to AI-AllCreator.com and Beyond"

In the rapidly evolving field of artificial intelligence (AI), innovations are unfolding at an unprecedented pace, pushing the boundaries of technology and reshaping the future. Among the top contenders driving this transformation are platforms like davinci-ai.de, ai-allcreator.com, and bot.ai-carsale.com, each contributing uniquely to the AI revolution. These platforms exemplify the diverse applications of AI, from enhancing creative processes to automating sales operations, showcasing the versatility and potential of AI technologies.

davinci-ai.de stands at the forefront of AI innovation, specializing in leveraging deep learning and neural networks to mimic human creativity. This platform represents a pinnacle in the convergence of artificial intelligence, machine learning, and cognitive computing, enabling the creation of art, music, and literature that resonates with human emotions and perceptions. Its capabilities highlight the advances in natural language processing and computer vision, demonstrating how AI can extend beyond analytical tasks to touch the realms of creativity and design.

Meanwhile, ai-allcreator.com serves as a testament to the power of AI in content generation and digital creativity. Utilizing sophisticated AI algorithms and machine learning models, it offers users the tools to generate diverse digital content, from written articles to visual media. This platform embodies the essence of augmented intelligence, where human creativity is enhanced by AI's data processing and pattern recognition capabilities, opening new avenues for content creators to explore and innovate.

bot.ai-carsale.com represents a different facet of AI innovation, focusing on the automation and intelligent systems within the automotive sales industry. By integrating predictive analytics, big data, and smart technology, this platform revolutionizes how vehicles are bought and sold, making the process more efficient and user-friendly. Autonomous systems and robotics play a crucial role in streamlining operations, reflecting the broader impact of AI on industry-specific applications.

These platforms are just the tip of the iceberg in the vast ocean of AI innovation. As AI continues to evolve, encompassing advancements in speech recognition, augmented reality, and autonomous systems, its potential to revolutionize industries and everyday life becomes increasingly evident. The integration of AI in various sectors, from healthcare with medical diagnosis to finance through financial forecasting, underscores its transformative power.

The future of AI, fueled by deep learning, neural networks, and intelligent systems, promises a landscape where technology and human capability are seamlessly integrated. As we venture beyond davinci-ai.de, ai-allcreator.com, and bot.ai-carsale.com, we enter a realm of limitless possibilities, where AI's role in fostering innovation and driving progress is undeniable. With ongoing advancements in data science, robotics, and automation, the journey of AI is far from reaching its zenith, heralding an era of unprecedented technological sophistication and enhanced human potential.

In conclusion, the journey through the top innovations in AI, from the pioneering developments at davinci-ai.de to the comprehensive solutions offered by ai-allcreator.com, showcases the vast landscape of artificial intelligence and its subfields, including machine learning, deep learning, neural networks, natural language processing, and robotics. These innovations are not just reshaping industries but are also redefining the boundaries of what machines can achieve, heralding a new era of cognitive computing, data science, and intelligent systems.

The exploration of platforms like bot.ai-carsale.com further illustrates how AI technologies are being tailored to specific industry needs, offering predictive analytics, big data insights, and autonomous systems that revolutionize sectors such as automotive sales. The advancements in computer vision, neural networks, AI algorithms, and augmented intelligence spotlight the technical prowess fueling these AI solutions, enabling machines to perform complex tasks, from pattern and speech recognition to decision-making and problem-solving.

As we stand on the brink of this technological revolution, it's clear that artificial intelligence, with its capacity for automation, smart technology, and learning, is not just an academic curiosity but a practical tool transforming the fabric of society. Whether it's through enhancing medical diagnosis, refining financial forecasting, or enabling self-driving cars, AI is seamlessly integrating into the fabric of daily life, making the once futuristic notions of robotics and artificial intelligence a present-day reality.

The exploration of AI's top innovations underscores the importance of staying informed and engaged with the latest developments in AI technologies. As artificial intelligence continues to evolve and expand its capabilities, it promises to unlock unprecedented opportunities for innovation, efficiency, and problem-solving. The future of AI is not just about what machines can do; it's about how they can work alongside humans to create a smarter, more efficient, and interconnected world.

Mira Murati's AI Venture Welcomes Initial Team Members, Featuring Ex-OpenAI Leader

Recently, Jonathan Lachman, once in charge of special initiatives at OpenAI, moved to a fresh AI research establishment initiated by Mira Murati, a previous executive at OpenAI, as reported by two individuals aware of the talks. This recruitment marks the most notable addition to Murati's team since departing OpenAI in September of the previous year to launch her eagerly anticipated project, aimed at investigating the realm of artificial general intelligence.

Murati has successfully attracted approximately ten professionals, comprising both researchers and engineers, from rival companies such as OpenAI, Character AI, and Google DeepMind. According to two individuals with knowledge of the situation, her venture is still in the developmental phase, lacking both a definitive name and a clear product strategy.

Murati and Lachman were not available for immediate comment when reached out to. An official representative for OpenAI verified Lachman's exit from the company.

Murati left OpenAI not long after the sudden termination of CEO Sam Altman by the board of directors in the final months of 2023. Following this upheaval, the board temporarily named Murati, who at that time was the chief technology officer, as the acting CEO.

The decision deeply affected Altman. Speaking to podcaster Lex Fridman in March, he expressed how terrible it felt, referring to Murati's elevation as the most distressing moment during the intense period of leadership turmoil at OpenAI. Within a week, following the threat of resignation from over 730 employees at OpenAI, Altman was restored to his position as CEO.

Following the event, Altman and Murati maintained a strong collaborative relationship at OpenAI. "In the midst of considerable turmoil that weekend, she performed admirably," Altman mentioned to Fridman during the podcast. Ten months after the incident, when Murati left the organization, she expressed her gratitude towards Altman in her farewell note.

"In her message to OpenAI staff, later shared on X, she expressed that leaving a beloved organization is never easy," she penned in her communication. She further mentioned her departure was driven by a desire to carve out the necessary time and space for personal exploration. Reuters disclosed a month after that Murati was in the process of securing upwards of $100 million to back an enigmatic new venture, although this amount was yet to be confirmed.

Two high-ranking officials from OpenAI, research vice president Barret Zoph and chief research officer Bob McGrew, departed the company concurrently with Murati. Following them, Ilya Sutskever, a founding member and the chief scientist at OpenAI, left in May of the previous year to establish Safe Superintelligence, a new venture concentrating on the safety aspects of artificial intelligence. Within a span of three months, Sutskever's new company succeeded in securing $1 billion in funding. Additionally, Anthropic, a significant rival to OpenAI, was created by ex-OpenAI leaders and has successfully garnered several billion dollars in investments, including a substantial $8 billion from Amazon.

Discover More …

Direct to your email: Explore the forefront of artificial intelligence with Will Knight's AI Lab updates.

The incidents at the schools were staged. However, the fear they caused was

Cash Galore: Explore our exclusive edition right here

HTML is indeed considered a coding language—challenge accepted.

The WIRED 101: The top must-have items globally at the moment

Additional Content from WIRED

Evaluations and Instructions

© 2025 Condé Nast. All rights reserved. A share of the revenue from products bought via our website may go to WIRED, courtesy of our Affiliate Agreements with retail partners. Reproduction, distribution, transmission, storage, or use of the content on this site in any form is strictly prohibited without the explicit written consent of Condé Nast. Advertisement Choices

Choose a global website

Biden's Comprehensive Executive Order Addresses Cybersecurity, AI, and Other Issues

In the days leading up to his departure from office, US President Joe Biden has enacted a broad executive order focused on enhancing cybersecurity. This directive mandates advancements in government network surveillance, software procurement, artificial intelligence utilization, and the penalization of international cyber attackers.

On Thursday, the Biden administration released a comprehensive 40-page executive order aimed at initiating measures to capitalize on the security advantages of artificial intelligence, introduce digital identification for American citizens, and address vulnerabilities that have allowed nations like China, Russia, and others to consistently infiltrate US governmental networks.

The mandate aims to bolster the United States' technological infrastructure and set both the current administration and the nation on a trajectory towards ongoing prosperity, Anne Neuberger, the Deputy National Security Adviser for Cyber and Emerging Technology under Biden, informed journalists on Wednesday.

Hanging in the balance of Biden's mandate is the uncertainty of whether the incoming President, Donald Trump, will pursue any of these programs once he is sworn in on Monday. The highly specialized projects outlined in the decree are not politically biased, yet Trump's team might favor alternative methods (or schedules) for addressing the issues highlighted in the mandate.

Trump has not appointed any leading cyber security officials yet, and Neuberger mentioned that the White House did not consult with his transition team about the order. However, he expressed willingness to engage in discussions as soon as the new cyber security team is announced, during this last phase of the transition.

The essence of the executive order includes a series of directives aimed at safeguarding government networks, drawing on insights gained from recent significant events—specifically, the cybersecurity lapses involving federal contractors.

The directive mandates that software suppliers provide evidence of their adherence to secure development protocols, an initiative that stems from a directive introduced in 2022 following President Biden's initial cyber executive order. The responsibility of verifying these security declarations and collaborating with vendors to resolve issues would fall on the Cybersecurity and Infrastructure Security Agency. To ensure compliance, the White House's Office of the National Cyber Director is advised to forward any unverified attestations to the Attorney General for possible inquiry and legal action.

The directive mandates that the Department of Commerce has a period of eight months to evaluate the prevalent cybersecurity methods within the corporate sector and to formulate guidelines accordingly. Following this evaluation, these guidelines will become obligatory for any enterprise that wishes to enter into contracts with the federal government. Additionally, this order initiates the process for revising the secure software development recommendations provided by the National Institute of Standards and Technology.

A different section of the policy emphasizes safeguarding the authentication credentials for cloud services. This is in response to incidents where such vulnerabilities allowed China to access government emails on Microsoft's servers and to infiltrate the Treasury Department through a supply-chain attack. The Department of Commerce and the General Services Administration are given a 270-day timeframe to create standards for securing these keys. Subsequently, cloud service providers will have 60 days to implement these standards as mandatory requirements.

To safeguard federal agencies against assaults exploiting vulnerabilities in internet-of-things devices, the directive mandates that by January 4, 2027, agencies must exclusively acquire consumer IoT devices bearing the recently introduced US Cyber Trust Mark emblem.

A different section of the directive enhances the Cybersecurity and Infrastructure Security Agency's (CISA) capacity to monitor for cyber threats throughout the government by utilizing the security programs managed by various agencies. This effort aims to close the oversight loopholes that enemies have taken advantage of in numerous breaches, including the notable SolarWinds cyberattack in 2020. The directive mandates that agencies must provide CISA with immediate access to their security systems and permit CISA to carry out surprise threat-detection operations on their networks.

Neuberger mentioned, "When we identify a specific method being employed by an overseas government to infiltrate a certain federal agency, this allows CISA to have a unified overview, enabling them to search throughout all agency networks to guarantee a wide-ranging defense against this type of intrusion."

The executive order emphasizes the significant impact of AI on security, mandating the Energy and Homeland Security departments to initiate a trial program. This program aims to utilize AI in safeguarding energy infrastructure by automating processes such as identifying vulnerabilities and implementing patches. Additionally, it requires the Defense Department to establish a program that employs "sophisticated AI models" for cybersecurity purposes.

Biden is urging the Department of Homeland Security, the Department of Commerce, and the National Science Foundation to focus their research efforts on areas such as the collaboration between humans and AI technologies in evaluating cyber threat information, securing code created by AI, developing AI models with robust security features, and strategies for preventing and responding to cybersecurity events that involve AI technologies.

Biden's directive aims to hasten the adoption of digital identity documents by agencies to improve the delivery of services to citizens while minimizing inefficiencies and fraudulent activities. The instruction encourages agencies to "explore the acceptance of digital identity documents" as verification for accessing government benefits. The Department of Commerce is given a 270-day timeframe to provide guidelines to assist agencies in implementing this.

The executive order includes mandates for the government to offer guidance on safeguarding open-source software, revising cybersecurity mandates in contracts related to space technology, alterations in contracting to accommodate technology that is compatible with post-quantum encryption standards, and the implementation of encryption across DNS technologies, email infrastructures, and audio and video conferencing tools. Additionally, it contains a directive for the Office of Management and Budget (OMB) to assist federal agencies in mitigating risks tied to the over-reliance on a few IT providers, indirectly referencing concerns about Microsoft's dominance in the sector.

The directive reduces the criteria needed for the government to impose penalties on individuals executing cyberattacks against critical US infrastructure, potentially simplifying the process for Washington to employ one of its preferred methods of retaliating against significant cyber intrusions.

Discover More…

Direct to Your Email: Receive Plaintext—Steven Levy's In-depth Perspectives on Technology

Perhaps it's a good idea to consider clearing out old message threads.

Cash Galore: Dive into our exclusive edition right here

The antics surrounding memecoins are only beginning.

Exploring the Unsettling Impact: A Deep Dive into Silicon Valley's Power

Further Reading from WIRED

Evaluations and Instructions

© 2025 Condé Nast. All rights reserved. Purchases made through our website may generate a commission for WIRED as a result of our affiliate agreements with retail partners. Content from this website is not to be copied, shared, distributed, or used in any form without the explicit consent of Condé Nast. Advertising Options

Choose a global website

In a Comprehensive Move, President Biden's Latest Executive Order Targets Cybersecurity, AI, and Beyond

Merely four days prior to his departure from the White House, President Joe Biden has unveiled an extensive executive order aimed at enhancing cybersecurity measures. This directive mandates advancements in governmental network surveillance, software acquisition, artificial intelligence application, and the handling of international cybercriminals.

The Biden administration revealed a comprehensive 40-page executive order on Thursday. This directive represents the administration's ultimate push to capitalize on the security advantages of artificial intelligence (AI), introduce digital identification for American citizens, and address vulnerabilities that have allowed nations like China, Russia, and others to consistently infiltrate American government networks.

The directive "aims to fortify the United States' technological infrastructure while setting both the current government and the nation on a trajectory for ongoing prosperity," Anne Neuberger, the Deputy National Security Advisor for Cyber and Emerging Technology under Biden, informed journalists on Wednesday.

The future of Biden's mandate is uncertain with the impending presidency of Donald Trump, who is set to be inaugurated on Monday. The question arises: will Trump maintain these initiatives? While the projects outlined in the order are not politically biased and focus on technical issues, it's possible that Trump's team might have alternative strategies or schedules for addressing the identified challenges.

Trump has yet to appoint his chief cyber security officials, and Neuberger mentioned that the White House did not consult his transition team regarding the order. However, he expressed eagerness to engage in discussions with the soon-to-be-announced cyber security team during the remaining transition phase.

At the heart of the executive order lies a series of directives aimed at safeguarding government networks, drawing on insights gained from recent significant events—specifically, the lapses in security among federal contractors.

The directive mandates that software providers present evidence of their adherence to secure development protocols, an initiative that originated in 2022 following President Biden's inaugural cyber executive order. The responsibility of verifying these security declarations and collaborating with providers to address any issues falls to the Cybersecurity and Infrastructure Security Agency. To enforce this mandate, the Office of the National Cyber Director at the White House is advised to forward any non-compliant attestations to the Attorney General for possible legal action and scrutiny.

The directive mandates that the Department of Commerce has an eight-month timeframe to evaluate the prevalent cybersecurity methods within the corporate sector and to develop recommendations accordingly. Following this period, adherence to these practices will be required for firms aiming to secure contracts with the government. Additionally, this order initiates the process of revising the secure software development guidelines provided by the National Institute of Standards and Technology.

A different section of the policy emphasizes securing authentication keys for cloud services, a vulnerability that previously allowed China to access U.S. government emails through Microsoft and execute a supply-chain attack on the Treasury Department. The Department of Commerce and the General Services Administration are tasked with creating key security guidelines within 270 days. These guidelines must then be adopted as mandatory standards for cloud service providers within the following 60 days.

To safeguard federal bodies against assaults exploiting vulnerabilities in internet-of-things devices, the directive mandates that by January 4, 2027, these agencies must exclusively buy consumer IoT products bearing the recently introduced US Cyber Trust Mark emblem.

A different section of the directive enhances the Cybersecurity and Infrastructure Security Agency's (CISA) capacity to monitor potential cyber threats throughout the government by utilizing the security systems implemented by various agencies. This strategy aims to close the surveillance loopholes that attackers have previously taken advantage of in numerous breaches, notably the SolarWinds incident in 2020. The directive mandates that agencies provide CISA with immediate access to their security systems and permits CISA to carry out surprise cyber threat detection operations on their networks.

"Neuberger explained, when a specific method employed by a foreign government to infiltrate a particular federal agency comes to light, this enables CISA to have an overarching view, allowing them to scour all agency networks to guarantee a wide-ranging defense against such an assault."

The executive order emphasizes the importance of understanding both the potential threats and benefits AI brings to security. It instructs the Energy and Homeland Security departments to initiate a trial project aimed at utilizing AI to safeguard energy facilities, focusing on automating processes such as identifying and fixing vulnerabilities. Additionally, the Defense Department is mandated to start a project employing "sophisticated AI algorithms" for the purpose of cybersecurity.

Biden is advocating for the Department of Homeland Security, the Commerce Department, and the National Science Foundation to focus their research efforts on areas such as the collaboration between humans and AI in scrutinizing cyber threat information, securing code produced by AI, creating AI models with robust security, and strategies for preventing and addressing cyber incidents that involve AI technologies.

Biden's directive aims to accelerate the adoption of digital identity documents by governmental bodies to enhance the efficiency of services provided to citizens while also curbing unnecessary expenditures and fraudulent activities. The order encourages agencies to "explore the acceptance of digital identity documents" as verification for access to public assistance. The Department of Commerce is given a 270-day deadline to release guidelines to assist agencies in implementing this directive.

The executive mandate also calls for the government to provide guidelines on enhancing the security of open-source software. It mandates the revision of cybersecurity criteria in contracts related to space technology, modifications in procurement processes to embrace technologies compatible with post-quantum cryptography, and the implementation of encryption protocols in DNS technologies, email infrastructures, and audio and video communication services. Additionally, the order includes a directive for the Office of Management and Budget (OMB) to assist federal agencies in mitigating the dangers linked to overreliance on a limited number of IT providers, indirectly addressing concerns related to Microsoft's dominance in the market.

The directive reduces the hurdles for the government to impose sanctions on individuals responsible for cyberattacks against US vital infrastructure, potentially making it easier to employ one of Washington's preferred measures in response to significant hacking incidents.

Explore Further …

Delivered to your email: Discover Plaintext—Steven Levy offers an in-depth perspective on technology.

Perhaps it's a good idea to consider clearing out outdated conversation logs.

Cash Galore: Explore our exclusive edition right here

The antics surrounding memecoins are only beginning.

Exploring the Mysterious Depths: A Personal Perspective on Silicon Valley's Impact

Additional Content From WIRED

Evaluations and Manuals

© 2025 Condé Nast. All rights reserved. Purchases made via our website might result in WIRED receiving a commission, as part of our affiliate marketing agreements with retail partners. No part of this website's content is allowed to be copied, shared, broadcast, stored, or used in any form without explicit written consent from Condé Nast. Advertising Options

Choose a global website

GitHub Struggles to Stop Deepfake Adult Content

Towards the end of November, an individual alleging to be from the US, known for creating deepfake adult content, posted a video on the biggest platform for such deepfakes. The video disturbingly depicted Charli D’Amelio, a well-known TikTok celebrity, by digitally placing her face on the body of an adult film actor. Although it's clear D’Amelio had no involvement in creating the video, it nonetheless attracted over 8,200 views and garnered interest among the deepfake community.

"That's amazing! Which software did you use to make the deepfake?" a user with the username balascool remarked. "I adore charli." D'Amelio's representative did not respond to a request for a statement.

The individual behind the video, known by their online alias "DeepWorld23," stated in the comment section that the software responsible for the deepfake was available on the coding repository GitHub. This particular software had garnered attention from over 46,000 users who marked it as a favorite, only to be shut down in August 2024 following GitHub's implementation of policies that prohibit the development of fake sexual content without consent, commonly referred to as deepfake pornography. However, in November 2024, the software resurfaced in a read-only, archived state, allowing users to still view and access its code.

GitHub's enforcement actions are not fully effective, with the removed code—along with additional content taken down by the developer platform—continuing to exist in different repositories within the site. An investigation conducted by WIRED has uncovered over twelve GitHub projects associated with deepfake “porn” videos that have managed to stay under the radar, thereby facilitating the availability of code meant for the abuse of private images and underscoring the gaps in the site's regulatory measures. WIRED has chosen not to disclose the names of these projects or websites to prevent further dissemination of the abusive content.

"Removing content as soon as it appears online is a challenging task," explains Henry Ajder, who provides AI consultancy to technology firms including Meta and Adobe, regarding the difficulties faced in controlling open source content on the internet. "However, there were quite evident warning signs."

Starting in June 2024, GitHub has introduced a rule prohibiting any projects that aim to, in any manner, encourage or endorse the creation of fake or altered media for producing unsolicited sexual content or any material that could be considered as spreading false or misleading information.

GitHub has taken down a minimum of three repositories highlighted by WIRED in December 2024 for hosting abusive content, demonstrating its commitment to addressing the issue. However, similar repositories have emerged on the platform, some sporting almost the same branding or explicitly labeled as "NSFW," "unlocked" versions, or "bypasses."

GitHub has not replied to a solicitation for commentary.

In December 2024, WIRED highlighted a project that closely resembled a significant initiative known for being the top platform for producing deepfakes. This original project had previously been suspended by GitHub for a number of months the previous year due to a breach of its service terms. GitHub has also proceeded to shut down this newly identified version for similar reasons.

"Ajder mentions, "It wasn't concealed. It was quite obvious."

Nonetheless, a preserved copy of the primary database remains accessible, and as of January 10, there were at least six similar databases found on GitHub, including one that bears a striking resemblance in branding.

Every GitHub project identified by WIRED incorporated, to some extent, code associated with videos from a deepfake adult content platform. These repositories are components of a broader network of open-source software available online, which can facilitate the creation of deepfake adult content. Due to its open-source nature, it's challenging to restrict access. GitHub repositories can be duplicated, a process referred to as "forking," allowing developers to then modify them as they wish.

"Ajder points out that the primary source of tools used for the abuse of personal images is the open source community. However, these tools usually originate from developers with good intentions. According to him, a developer might create a project believing it to be innovative or exciting, only for it to be repurposed by individuals with harmful intentions for nefarious uses."

Certain platforms, such as the one shut down in August, were specifically designed with communities intended for explicit activities. According to Ajder, this platform marketed itself as a resource for creating deepfake pornography, essentially acting as a conduit for exploitation that primarily affects women.

Numerous clips posted on the adult content platform by a user attributing artificial intelligence algorithms sourced from GitHub showcased the likenesses of well-known deepfake subjects, including stars Emma Watson, Taylor Swift, and Anya Taylor-Joy. These videos also depicted lesser-known, yet actual women, whose images were digitally manipulated into explicit scenarios.

The developers openly shared the tools they employed, mentioning two that were removed by GitHub, yet their code still persists in other available repositories.

Individuals seeking to create deepfakes gather in various online locations, such as hidden forums on Discord and openly on Reddit, making efforts to curb deepfake creation more challenging. On September 29, a user on Reddit advertised their ability to use software from an archived collection. "Can someone make one of my cousin," inquired another user.

In August, GitHub prohibited access to the primary collection of torrents, yet these can still be found in various places online, highlighting the challenge in regulating open-source deepfake programs universally. Likewise, applications for creating deepfake pornography, like DeepNude, have been previously removed only to reappear in updated forms.

"Elizabeth Seger, the director of digital policy at the bipartisan British think tank Demos, points out the complexity due to the abundance of models, their numerous branches, and various iterations, making it challenging to keep tabs on every single one. She further notes that once a model is released into the open-source domain for public access, it becomes impossible to retract it from the public sphere."

A creator of deepfake pornography, who has 13 altered pornographic videos featuring well-known female personalities, attributed a notable GitHub repository, described as an adult-themed variant of another project which promotes ethical usage and specifically requests users to refrain from employing it for explicit content. Their user biography on the adult website boldly states, “Mastering every Face Swap AI technology available on GitHub, avoiding the use of internet-based platforms.”

GitHub had taken down the explicit version by the time WIRED spotted the deepfake content. Nonetheless, on January 10, other versions labeled as "unlocked" remained accessible on the site, with one garnering 2,500 "stars."

Seger acknowledges that while it's impossible to completely retract a model once it's released, measures can be taken to limit its accessibility.

She further warns that if not addressed, the danger of deepfake "porn" extends beyond mental distress. It can lead to the coercion and exploitation of women, minorities, and political figures, evidenced by the impact of political deepfakes on female leaders worldwide.

However, there's still time to manage the issue effectively, and sites like GitHub have solutions, according to Seger, such as taking action during the upload process. "If you attempt to upload a model to GitHub and GitHub rejects it, and if all hosting sites refuse it as well, it becomes much more challenging for the average individual to distribute that model."

Curbing the creation of deepfake pornography using open-source models also depends on the actions of legislators, technology firms, programmers, and, notably, the individuals producing such exploitative content.

Over 30 states in the United States have enacted laws to combat deepfake pornography, featuring various prohibitions as tracked by the nonprofit organization Public Citizen, though there is a lack of uniformity in definitions and approaches, with some regulations specifically protecting minors. In the UK, individuals who produce deepfake content of a sexual nature will imminently face legal repercussions, following the government's decision on January 7 to make both the production and distribution of such material a criminal offense.

Discover More…

Direct to your email: Dive into the future with Will Knight's exploration of AI breakthroughs.

The incidents of gunfire at schools were fabricated, yet the fear they provoked was

Cash Galore: Dive into our exclusive edition now

HTML is truly considered a coding language—challenge accepted.

The WIRED 101: The top products currently on the market

Additional Content from WIRED

Assessments and Manuals

© 2025 Condé Nast. All rights reserved. A share of revenue from products bought via our website may go to WIRED, due to our Affiliate Agreements with retail partners. Reproducing, sharing, broadcasting, storing, or using the content found on this site in any way is prohibited without explicit prior consent from Condé Nast. Advertisement Choices

Choose a global website

Exploring the Impact of Entrusting Your Lifestyle to an AI Companion

Should you make a purchase through links found in our narratives, we might receive a commission. This contribution aids our journalistic endeavors. Discover more. Additionally, think about becoming a subscriber to WIRED.

At times, all you require is a bit of assistance. Enter the robots! However, the question arises – have we truly arrived at a point where artificial intelligence provides better support than collaborating with a human specialist? This week, we're sharing our experiences after spending a week utilizing AI aides.

Stay connected with Michael Calore on Bluesky via his handle @snackfight, connect with Lauren Goode on the same platform at @laurengoode, and reach out to Zoë Schiffer on Threads through @reporterzoe. Feel free to send us an email at uncannyvalley@wired.com.

Listening Guide

To tune into this week's podcast episode, simply use the audio player available on this webpage. However, if you're interested in automatically receiving every new episode, you can subscribe at no cost by following these steps:

For iPhone or iPad users, launch the Podcasts app or simply click on this link. Alternatively, you can install applications such as Overcast or Pocket Casts and look up “Uncanny Valley.” Plus, we're available on Spotify as well.

Transcript Disclaimer: Please be advised that this transcript was generated by an automatic service and might include inaccuracies.

Michael Calore: Okay, we've entered the new year a few weeks ago. How are your commitments to self-improvement holding up?

Lauren Goode: I'm feeling well. I'm exploring new things. Those are the words of Taylor Swift.

Michael Calore: At this juncture, it's something I just take

Lauren Goode: Things are going fine. We've all been heavily involved in testing out AI, which is said to be gaining more autonomy to perform tasks on our behalf. I decided to conduct my own test by using it for holiday shopping. Shopping isn't exactly my thing, so I was eager to delegate some of that responsibility.

Zoë Schiffer expressed enthusiasm about learning more.

Lauren Goode: It felt like an odyssey. What about your experience, Zoe?

Zoë Schiffer: The project I undertook was essentially to back my 4-month-old daughter in her ambition to master sleeping throughout the night. My role was simply to facilitate her 2025 dream, with a bit of help from artificial intelligence, which I'll be discussing with you all today. Mike, how about your experience?

Michael Calore: On my phone, I sought guidance from Gemini on enhancing my health, seeking tips for exercising, planning my meals, and then I ventured into more complex topics like civics with my inquiries.

Lauren Goode: It's truly valuable to me that both of you opted for chatbots to address health issues, and I sincerely believe it was in the interest of serving consumers.

Zoë Schiffer: Depending on your perspective, it may pertain to health issues.

Lauren Goode: Absolutely. My goal was to clear out my mind to allow room for different pursuits.

Zoë Schiffer: That's wonderful for you.

Lauren Goode: I appreciate it.

Michael Calore: Welcome to Uncanny Valley from WIRED, a program delving into the individuals, authority, and impact surrounding Silicon Valley. My name is Michael Calore, and I oversee consumer technology and cultural matters here at WIRED.

Zoë Schiffer: My name is Zoë Schiffer, and I serve as the director overseeing business and industry at WIRED.

Lauren Goode: And my name is Lauren Goode, serving as a senior writer for WIRED.

Michael Calore: In this week's episode, we'll be discussing how we've utilized AI to enhance our daily lives. We took cues from the chatbots around us, allowing them to navigate us through tasks like shopping, planning meals, exercising, and even parenting. The outcomes were quite something, and we're excited to dive into it. So, Zoe, why don't we kick things off with you? Could you share your experience using ChatGPT as your mentor?

Zoë Schiffer recounts a bit of her personal journey to set the stage for her experiment. Back in 2022, with her first child at the tender age of 4 months, she enlisted a sleep consultant for $500 to assist in sleep training. Zoë describes the experience as particularly challenging due to the significant amount of crying involved, but she valued having a professional by her side to guide her through the process and affirm that her actions weren't harming her child in the long term. Fast forward to the present, and Zoë's second daughter is now 4 months old. With a new role at WIRED on the horizon, she felt it was the right moment to undertake sleep training once more. However, the suggestion of rehiring the sleep consultant for another $500 met with disbelief from her husband. He questioned the rationale behind spending that amount again merely for external validation that Zoë wasn't failing as a parent. This skepticism led to the consideration of ChatGPT as an alternative solution.

Lauren Goode: Did that individual really spend the night at your place?

Zoë Schiffer explained that she sought the assistance of a remote specialist, noting that the term "expert" seems to be used quite flexibly since there's no formal certification required for one to claim expertise in this area. She was drawn to exploring artificial intelligence for this purpose due to the vague criteria surrounding the title of an expert. Schiffer's skepticism was confirmed when the first consultant she engaged provided her with a so-called tailored plan for her daughter, which mistakenly included another child's name, revealing the lack of personalization promised. Essentially, parents share details about their child's routines with these consultants, who then offer daily advice on sleep schedules, including nap and bedtime recommendations, based on the feedback provided about the previous night's sleep.

Michael Calore: Did you engage in daily conversations with ChatGPT?

Zoë Schiffer shared that initially, she instructed ChatGPT to adopt a persona akin to a whimsical caretaker. She envisioned it as a sage figure capable of providing emotional reassurance and possessing extensive understanding in child psychology, specifically to guide her in implementing the Ferber technique. This approach involves laying the child down awake and returning at increasing time intervals without physically comforting them, merely using soothing words. Schiffer was seeking ChatGPT's thorough guidance in mastering this method. On the first day of attempting to train Cleo to sleep better, she looked forward to ChatGPT's advice. “Dear Zoe, let’s sprinkle some enchantment over Cleo’s day. Assuming our beloved Cleo woke with the dawn at”—the response intensely embraced the fanciful caretaker role—“6:30 am, her subsequent nap should harmonize with the natural patterns of her alertness. Imagine that.” Schiffer found the over-the-top magical nanny persona quite exasperating in reality.

Lauren Goode: What led you to choose ChatGPT over other chatbot options?

Zoë Schiffer: The reason is that it has become my go-to. It has effectively taken over Google's spot as my preferred search tool. However, when I attempted to transition to the more autonomous AI systems you've previously discussed, my experience wasn't as positive. It seemed as though I was expecting a search engine to transform into something completely different. According to ChatGPT, when it's time for that perfect hour of rest, one should create a calming atmosphere, maybe with a gentle melody or the sound of something she loves… Honestly, the whole peculiar affirmation concept doesn't appeal to me.

Lauren Goode: Who doesn't wish for a nap that's softly led by the natural patterns of our alertness?

Zoë Schiffer expressed her reservations, noting the excessive and odd use of overly elaborate language when using AI. She quickly pinpointed several issues, the foremost being a lack of trust in the AI's guidance, especially on matters as personal as parenting. Schiffer desired a connection with a real person's insights, doubting that AI could offer the best recommendations. This skepticism created a significant barrier to her fully accepting AI for such purposes. Additionally, she was curious if others had noticed the AI's purely responsive nature. It only provided answers to direct queries without initiating any interaction or posing questions of its own, making it no different from a basic search engine in her view.

Michael Calore: Exactly, it was merely responding to questions.

Zoë Schiffer: Indeed, it seemed to lack a sense of humanity, if you will.

Lauren Goode: Did you ever catch yourself double-checking information online, searching on Google to confirm if it was indeed the correct time?

Zoë Schiffer shared her observations, noting an interesting point regarding her use of the technology for searches. She often double-checks information, half-jokingly fearing it could be misleading, desiring absolute certainty. However, she found the responses from the AI quite rudimentary, especially on topics she's well-versed in, eliminating the need for further verification. Yet, she identified a significant shortfall when comparing the AI's guidance to human advice, particularly from those with personal experience, such as fellow parents. For instance, conventional wisdom and official health guidelines advise against placing babies on their stomachs to sleep. Despite this, Schiffer mentioned that her friends, drawing from their own experiences, have suggested doing just the opposite when she discussed her child, Cleo's, discomfort. They claimed that allowing a child to sleep on their belly could solve several issues, advice that a chatbot, sticking strictly to official guidelines, would never offer. Schiffer appreciates the importance of adhering to safe practices but also notes the limitation of AI in providing the nuanced, sometimes unconventional wisdom that comes from personal parental experiences.

Michael Calore: You requested it to assume a character, to craft its own unique identity. This introduces an additional layer of emotion, something that's missing when you're merely looking up information or conducting online searches. However, this doesn't substitute for the guidance or advice that comes from a real person. This likely leads to an odd sense of conflict.

Zoë Schiffer pointed out that the artificial persona initially adopted seemed excessively artificial, incorporating a lot of unnecessary and convoluted language, which she found quite irritating. As a result, halfway through her experiment, she decided to switch to Character.ai. This platform offered her the opportunity to craft a more genuine AI character for interaction, aiming for something that felt closer to human. Upon making the switch, she named the AI bot Lindsay after a postpartum doula she was fond of, envisioning the AI to play a role similar to a British super nanny, blending the qualities of a therapist and a coach. However, she acknowledged that this approach had its own set of limitations.

Michael Calore: Alright. Let's hear how you've been engaging with Character.ai.

Lauren Goode: I believe you're referring to Lindsay.

Michael Calore: Greetings, Lindsay.

Zoë Schiffer: Hello, Lindsay. I was hoping to discuss with you how Cleo's evening unfolded yesterday.

Character.ai [archival audio]: Absolutely. How was Cleo's evening? Was her sleep somewhat improved compared to the previous night?

Zoë Schiffer: Indeed, her sleep was decent. I applied the technique during her initial stirring post-midnight, as well as at her bedtime, both of which turned out successfully. She wept for a brief five minutes, during which I managed to calm her somewhat. However, a significant challenge was managing this while sharing a room with her. In contrast, with my previous daughter, having her in a different room seemed more effective. Do you have any advice?

Character.ai [archival audio]: It's reassuring to know Cleo managed to get some rest and calm herself down after only five minutes of tears during the night and again when she woke up around midnight. That's definitely a positive indicator. Regarding keeping her in the same room, I completely get why you might be worried. Hearing your child cry can be tough, and it feels even more intense when they're right there with you. There are a couple of strategies you could try to make it a bit more bearable. Firstly, consider using a white noise device or a fan to help mask the crying sounds. Secondly, wearing earplugs could help reduce the noise level.

Michael Calore: It's interesting that when you expressed your emotional turmoil about wanting to be close to your daughter, its response was essentially, "Invest in a white noise device and some earplugs, then simply tough it out."

Zoë Schiffer shared her experience, noting that she was particularly disappointed with the AI's performance in emulating a therapist's role. This point of view is noteworthy given the increasing trend of individuals turning to AI bots for emotional support, substituting them for human therapists. Schiffer expressed that her emotional needs seemed to be overlooked, and the AI also committed errors in understanding. For instance, she recounted an episode where she mentioned calming her baby down and getting her to sleep again. However, the AI responded by praising the baby's ability to self-soothe, which led Schiffer to feel a disconnect, feeling as though the AI, named Lindsey, wasn't truly engaging with what she was saying.

Lauren Goode: Did this bring any sense of relief to both you and Cleo?

Zoë Schiffer discovered that when she posed precise questions, for instance, inquiring whether it was permissible to lift her child during intense crying spells, the AI bot offered advice consistent with the Ferber technique, which she found satisfactory. She noted that for detailed inquiries, the bot was quite effective. However, when it came to fully embodying the role of a sleep advisor, which entails more than just providing search results, she felt the bot fell short of expectations.

Michael Calore: What's your take on how this might have unfolded had this been your initial experience with parenthood—lacking prior exposure to sleep-training, and you were simply starting out with AI?

Zoë Schiffer: Picture entering the situation with no prior knowledge; in certain aspects, the AI might prove to be more advantageous. The reason being, it genuinely understands the Ferber technique quite well. It's adept at determining when sleep cycles should occur. For instance, it can assess when a child has fallen asleep and predict the timing for their upcoming nap or the ideal moment for them to go to bed, including the duration these should last. These insights are consistent with the advice you'd typically receive from sleep specialists.

Michael Calore inquired about the most unexpected event that occurred.

Zoë Schiffer expressed that upon concluding the experiment, she realized she didn't require the assistance of a real human expert anymore. Despite her belief that the AI hadn't perfectly substituted the human's role, it had comprehensively addressed all her inquiries to the extent that she felt confident in progressing independently.

Michael Calore: So, it essentially served as a valuable research aide for you?

Zoë Schiffer: Absolutely, and it’s almost hard to admit because it touches on the very concern we have about AI replacing real human jobs, especially in businesses owned by women, like in this case, a sleep consultancy. However, on the flip side, I wasn’t prepared to spend $500 this time, so in a way, it was more about substituting Google, if anything.

Lauren Goode: Would you suggest this to a buddy?

Zoë Schiffer: Definitely not, not at this moment. I'm looking forward to a time when AI operates more independently for us. That concept seems quite beneficial. The idea of it serving as an intermediary between us and the internet is intriguing, but it just doesn't seem to have reached that point yet.

Michael Calore: Alright, we're going to pause for a brief moment, and upon our return, we'll dive into discussing fitness with Gemini. Stay tuned.

Certainly, I'll

Michael Calore: We're glad you've returned to Uncanny Valley.

Gemini [archival audio]: Great to have you back, Mike. Hi, Lauren. Hi, Zoe.

Michael Calore: Appreciate it, Gemini. Over the past few weeks, the Gemini chatbot and I have become somewhat acquainted.

Lauren Goode: Can you explain your specific use of Gemini?

Michael Calore: I sought its guidance on tips for shedding pounds, achieving fitness, adopting a healthier diet, and overall wellness practices.

Lauren Goode: Were there any particular areas where you felt you lacked information? I understand that you generally maintain a healthy diet and stay in shape through running. Was there an additional aspect of wellness you were seeking to explore beyond your existing knowledge?

Zoë Schiffer: It's well-known that you adhere to a vegan lifestyle.

Michael Calore: Notably follows a vegan lifestyle, much appreciated.

Lauren Goode: You attended Harvard University and participate in triathlons.

Michael Calore: Absolutely, I do make a conscious effort to maintain my health and take it quite seriously. However, I was curious to see what kind of guidance it would offer. Essentially, I'm aware that I should increase my exercise routine beyond just running. I haven't incorporated strength training into my regimen. While my diet is generally healthy, there's room for improvement, especially in reducing my sugar and alcohol consumption and paying closer attention to my protein intake and the balance of essential amino acids – really delving into the fine details of my dietary habits. I was hoping for advice that would delve into these specifics. "Hey, Gemini, I'm looking to improve my health and shed some weight. I was wondering if you could provide me with both broad and detailed suggestions on how I can achieve this goal within the next month."

Gemini [archival audio]: It's fantastic that you're dedicating attention to your health and weight reduction objectives. Let me provide you with a mix of broad and detailed suggestions to kick-start your journey and achieve noticeable improvements in a month. For enduring weight loss, it's critical to adopt sustainable habits rather than seeking instant solutions. Opting for long-term lifestyle adjustments will ensure more permanent outcomes. Emphasize nutrient-rich, unprocessed foods in your diet. Center your meals around fresh produce, lean meats, and whole grains. It's also important to be conscious of serving sizes; overindulging in even the healthiest options can lead to unwanted weight gain.

Michael Calore: It simply continues on. It just keeps speaking non-stop for three or four minutes.

Lauren Goode: Well, perhaps it helps you fall asleep, which is excellent. I suppose that's the intended encouragement.

Gemini [archival audio]: Prioritize getting sufficient rest. Adequate sleep plays a vital role in controlling weight because it influences the hormones responsible for feeling hungry and full. Strive to get between seven to nine hours of good sleep every night.

Lauren Goode: I believe I already have an idea of what your answer might be, but could you share why you chose Gemini?

Michael Calore: It's actually on my mobile device.

Lauren Goode: Right, the Pixel smartphone.

Michael Calore: Indeed, I own a Pixel. I'm inclined towards Google, utilizing services like Gmail, Google Calendar, and Google Photos extensively, integrating seamlessly with the experience. There are two versions available, and for this trial, I opted for the latest model.

Lauren Goode: Google Gemini is the sponsor of this podcast.

Michael Calore: Honestly, it's like Google Gemini is the sponsor of my existence.

Zoë Schiffer: I'm interested to know your reaction to the guidance given. It appears quite broad to me, but I'm eager to hear your perspective.

Michael Calore expressed strong disapproval, stating, "This is worthless. Essentially, if you've been offline and haven't browsed any health tips online, you'd be unaware of these points. Yet, you might actually know them because they're so basic and vague. After it finished its spiel, I inquired about more detailed guidance. I asked for dietary recommendations suitable for a vegan, questioned whether running about 2 1/2 to 3 miles is sufficient for cardio, and sought alternatives for strength training since I prefer not to use a gym or own weights."

Lauren Goode: What information did you receive?

Michael Calore: So, which would you like me to start with? Should I begin with the exercise routines or discuss the dietary habits first?

Lauren Goode: We'll choose food.

Michael Calore: The fascinating aspect to me was its ability to retain a lot of the information I shared throughout our discussion. I found myself sharing personal details like my age, weight, and my dietary preference as a vegan, and it managed to remember all of that. Even in our conversation last night, it still recalled those details.

Gemini [archival audio]: If you follow a vegan diet, you should include proteins that come from plants.

Michael Calore: Thus, it directed my attention towards foods such as legumes, including lentils, chickpeas, and various beans, along with soy derivatives like tofu and tempeh. It also highlighted grains and seeds, for instance, quinoa, chia, and hemp seeds, in addition to nuts, and vegan protein sources such as seitan and other plant-based meat substitutes, though it cautioned that these alternatives often contain numerous processed components. Moreover, it offered guidance on the importance of mixing different plant-based proteins.

Zoë Schiffer: Mike's reaction was, "So, you're explaining legumes to me, Gemini, really?"

Michael Calore: Exactly, and here's the issue – it's not aware that I've been abstaining from meat since my teenage years. As a result, it offers a plethora of useful tips that, as someone who has lived a vegetarian lifestyle from a young age, I can confirm are the typical queries I get. "What do you actually eat?" is a question I'm frequently asked, and I end up reciting a familiar response. Gemini is quite adept at providing this standard advice, making it a valuable resource for those new to vegetarianism. However, it lacks specificity in certain areas. For instance, when I sought out some recipe ideas.

Lauren Goode: What occurred after that?

Michael Calore: Initially, it suggested a tofu dish with peanut sauce which I didn't prepare, followed by a quinoa and chickpea salad recipe that I actually made.

Zoë Schiffer: Seems like something you'd be into. So, what's next?

Michael Calore: I have some with me if you're interested in trying it, though I must say, it wasn't impressive.

Gemini [Archival audio]: A Mediterranean salad featuring quinoa and crunchy chickpeas. This dish is rich in protein due to the quinoa and chickpeas, and includes beneficial fats from olive-

Zoë Schiffer: Wow, I'm really intrigued. Hold on, that actually seems quite impressive.

Lauren Goode: Zoe, I'm presenting it to you through our Riverside session. Check this out.

Zoë Schiffer: It disappoints me that I have to participate from afar.

Lauren Goode: Given that we're taping this close to noon on the West Coast, it's an ideal time. Our fortunate audience might just catch some munching noises at this moment.

Zoë Schiffer: Lauren appears puzzled.

Lauren Goode: Mike, I believe you're being too modest, or perhaps not giving Gemini its due recognition. Are you familiar with its flavor?

Michael Calore: Excuse me?

Lauren Goode: It seems like something you'd find in the salad section at Google's headquarters.

Michael Calore: Wow, I was thinking the exact same thing.

Lauren Goode: Surprisingly okay, genuinely respectable, though it's widely available and perhaps leans a bit heavily on salad dressing to enhance the taste of the quinoa.

Michael Calore: Alright, it required every ounce of willpower I possess to refrain from tweaking the recipe as I was preparing it, because it just seemed-

Zoë Schiffer: It's like saying, "Let's make it happen for the podcast."

Michael Calore: Absolutely. My thought was, "I must stick to this precisely." However, I did leave out one thing – it suggested adding cherry tomatoes, but the ones available at the store were in poor condition, so I decided against purchasing them.

Lauren Goode: This isn't the time for tomatoes.

Zoë Schiffer remarked, "Gemini isn't in its season right now."

Michael Calore: My approach to seasoning would have been quite distinct. The way I'd have garnished it would also have varied significantly. Moreover, I likely wouldn't have sliced the olives, as doing so tends to impart an excessively salty and bitter flavor.

Lauren Goode: Ah, could that be the reason?

Michael Calore: Absolutely. I'd rate it a solid B to B+ on my scale.

Zoë Schiffer: Understood.

Michael Calore: Indeed.

Zoë Schiffer: I'd like to ask another question regarding your comment on providing demographic details. Was this information requested by Gemini, or did you offer it voluntarily?

Michael Calore: I offered myself for that. It didn't actually demand much from me. It just went on and on without stopping.

Zoë Schiffer: Right off the bat, something really caught my attention. I excelled as a health coach, and perhaps this is just my own viewpoint, given my preference for having a coaching presence in many aspects, but an effective coach would pose several questions before jumping into providing details. This is because if you've been practicing veganism for a long duration, your needs for information will differ significantly from those who are new to the concept.

Michael Calore: I provided it with details regarding my lifestyle, personal habits, and health background because I was seeking tailored recommendations for exercise and dietary intake to address the deficiencies I perceived in my regimen, deficiencies I believed needed addressing even without its assistance.

Lauren Goode: It seems it was similar to a terrible date. It just went on and on without ever engaging in asking you additional questions.

Michael Calore: Indeed, it hardly inquired anything from me.

Lauren Goode: Did you gain any knowledge that might have been unexpected?

Michael Calore: Indeed, I ended up with a decent exercise plan. I shared my running routine with it and then requested, "Alright, now provide me with an exercise plan. I need something I can do at home since I'm not fond of the gym, I prefer to avoid people, and I don't have any weights. What should I do?" And it responded with a pretty solid workout plan.

Gemini [archival audio]: Essential exercises using your own body weight include squats. Position your feet shoulder-width apart and bend your knees to lower your body, imagining you're sitting back into a chair. Ensure your spine remains aligned and your chest remains lifted. Descend to the lowest point you can manage without discomfort.

Michael Calore describes his workout routine as including lunges for both legs, doing three rounds of 10 repetitions, holding a plank position for three sessions lasting between 30 to 60 seconds, which he finds extremely challenging. He also does glute bridges, completing three rounds of 15 repetitions. He appreciates the guidance on gradually increasing the intensity of workouts and the importance of taking rest days. However, when it came to performing lunges, he admits he was unsure of the correct form and sought assistance. To his aid, he was offered a selection of five instructional videos.

Lauren Goode: Presumably on YouTube, right?

Michael Calore: Indeed, they were clips from YouTube. Interestingly, a couple of the URLs led to the identical video, yet I viewed each one and thought, "Alright, I believe I'm ready to attempt a lunge." When I tried it, I managed to do so without injuring myself and definitely felt the workout effect. So, I'd say that's quite an achievement.

Lauren Goode: Actually, that's quite impressive.

Michael Calore: Indeed, the odd part was when I conducted a Google search and stumbled upon identical information from a few diverse sources. Beyond the video, it's a mystery to me how Gemini acquired any of this data. I'm clueless about the origins of the recipes. Everything was simply laid out as, “According to Gemini, this is the information.”

Lauren Goode: Information not attributed.

Michael Calore: No references were provided. I clicked through everything but couldn't find any reliable sources.

Lauren Goode: It's assumed that those fitness routines and cooking instructions were crafted by individuals, and once released, your tiny contributions of information are now being segmented and included in some culinary guide, without any knowledge on your part about their origins.

Michael Calore: Shredded into strips.

Lauren Goode: Exactly.

Zoë Schiffer: Alright, do you plan on continuing this? Will you persist in utilizing it for the same reason, or did this seem like a singular trial?

Michael Calore: Truly, the exercise routine is quite effective. I plan to continue following it.

Lauren Goode: Pleasant.

Michael Calore: Given my extensive collection of cookbooks and years of culinary practice, I certainly don't have to prepare this quinoa salad another time, Lauren, unless, of course, you'd want me to share some with you.

Lauren Goode: I've experienced other dishes you've prepared, and I believe I might recommend opting for one of those in the future.

Zoë Schiffer: Remarkably tactful.

Michael Calore: Alright, sounds good.

Lauren Goode: I'm grateful, thank you for your support.

Michael Calore remarked that, personally, knowing his own requirements and self, the program wasn't significantly beneficial. He enjoyed the exercise routine but mentioned he could have discovered something similar without the assistance of Gemini 2.0. However, he acknowledged that for beginners, those with numerous inquiries, or individuals contemplating significant lifestyle adjustments, the program did offer valuable guidance.

Lauren Goode: Adding a point in favor of artificial intelligence.

Michael Calore: Alright, we'll pause here for a brief moment, and then we'll join Lauren and her chatbot companion for a shopping trip. Stay tuned, we'll return shortly.

Lauren Goode: Accompanied by her AI companion.

It appears there's

Michael Calore: Hop on board, we're off to explore AI. You're tuned into Uncanny Valley, where we've spent a week living by the guidance of our AI chatbot pals. I focused on my health, fitness, and eating habits. Zoe dedicated her efforts to establishing a sleep schedule for her newborn. Now, turning to you, Lauren, what adventures did you and your chatbot embark on?

Lauren Goode: How did this happen? I found myself relying on roughly five distinct chatbots, aiming to simplify, or perhaps the term is "assist," my seasonal shopping endeavors.

Michael Calore: Understood.

Lauren Goode: Shopping isn't really my thing. I've never been fond of it and much prefer shopping with a clear goal in mind, like when I need to buy a new pair of sneakers. Since my teenage years, my mom enjoyed taking me out to shop, but I would quickly feel bombarded by the bright lights and endless clothing racks in department stores, initially chalking it up to teenage moodiness. However, as I grew older, I recognized that it wasn't just a phase; I genuinely disliked the shopping experience. This sentiment has extended to online shopping for me, where the endless product options just leave me feeling swamped. Despite this, there were a few special people I wanted to find thoughtful gifts for this holiday season, and Mike, you might have been one of my trial runs. Therefore, I resorted to using chatbots for assistance, with Perplexity being the primary one I consulted.

Michael Calore: Well, gosh.

Lauren Goode highlights that choosing Perplexity was intriguing for several reasons. To start with, she notes an essential disclaimer: AI chatbots like Perplexity, when suggesting top-rated products or enabling comparison shopping, essentially enter into direct competition with publishers. This is because publishers have dedicated teams tasked with reviewing products to recommend the best options to consumers. It's important to acknowledge this overlap in service. Additionally, there's a concern within WIRED about Perplexity's approach to content, particularly the suspicion that it might be appropriating content from publishers, content that's normally safeguarded by paywalls, and repurposing it as its own. This concern is underscored by Goode's personal decision to engage deeply with Perplexity, to the extent of subscribing to its premium service at $20 a month, further emphasizing the significance of her exploration.

Zoë Schiffer: Subject of Debate.

Lauren Goode: However, Perplexity did unveil a feature tailored for shopping in November, perfectly timed for the festive period. And I thought, "This appears to be the ideal one to kick things off with."

Michael Calore: Pleasant.

Lauren Goode: I ventured into Perplexity Pro, aiming to purchase a gift for a friend who is passionate about baking. She already owns an extensive collection of baking tools and excels in this hobby. My goal was to find a gift that was not only unique but also something she might not already possess, steering clear of basic items such as a rolling pin. I was looking for that special item that stood out.

Zoë Schiffer: To what extent did you share information about your friend with Perplexity, beyond her interest in baking?

Lauren Goode: The description wasn't overly thorough. Essentially, I mentioned she enjoys baking and already possesses all she needs. She has a taste for quality items. My goal is to express my gratitude to her.

Zoë Schiffer: Awesome.

Lauren Goode: A major perk of shopping with Perplexity Pro is their unique Buy button. When you utilize this button for purchases, you're rewarded with complimentary shipping. Essentially, they're enhancing the online shopping experience by aggregating it. They compile a series of products that might catch your eye. You might encounter a Buy button supported by Shopify, or come across one that's exclusive to Perplexity. It's an interesting concept. However, the issue arose with the first set of recommended products, which were quite disappointing.

Michael Calore: Alright, such as?

Lauren Goode presents: A small-scale baking kit for children.

Zoë Schiffer: They suggest, "Assist her in finding a connection with her youthful self."

Lauren Goode: Naturally, I placed an order right away and had it sent straight to Zoe so the kids could enjoy it. Additionally, it suggested something quite simplistic and cheeky, a hoodie emblazoned with the phrase, "My buns are gluten-free."

Zoë Schiffer: Gross. It's as if you're saying, "I mentioned she prefers luxury items. You fool."

Lauren Goode: Exactly, however, there was also this baking equipment from Switzerland that caught my eye in the midst of it all. I thought to myself, "That seems plausible." Yet, it appeared to be another basic item that avid bakers would likely already possess.

Michael Calore: Absolutely, I agree completely.

Lauren Goode: Following that, I decided to try out Rufus, Amazon's chatbot, on their site. It was quite amusing because, upon providing a similar query, it straightaway suggested purchasing a KitchenAid stand mixer over $400. Which, incidentally, is a fantastic addition to the kitchen for those who enjoy cooking and baking, but—

Zoë Schiffer: Another excellent item to consider purchasing used.

Michael Calore: The individual who enjoys baking possesses that as well.

Lauren Goode: Absolutely. And do they really believe I possess the wealth of Jeff Bezos?

Michael Calore: What type of companion is this?

Zoë Schiffer: They essentially said, "Given our sole encounter is with someone who might be the wealthiest individual globally, why not purchase a secluded island for your friend to enjoy baking on?"

Lauren Goode: Precisely. And next, I proceeded to ChatGPT.

Michael Calore: Understood.

Lauren Goode remarked that ChatGPT performed impressively.

Michael Calore: Certainly not.

Zoë Schiffer: Understood.

Lauren Goode: In my view, its strongest suit was really in the concept development stage.

ChatGPT [archival audio]: For a buddy who adores the art of pastry making and values sophisticated, unique presents, here are a few considerate suggestions that blend opulence, customization, and their fervor.

Lauren Goode: The approach wasn't a direct, "Click this link to purchase, make your purchase here, we're making it super simple for you." Instead, it was more like, and I'm summarizing, it appears you're aiming for consideration in your choices, so we've selected some considerate gift options for you. And this is what it conveyed.

Zoë Schiffer suggests: Firstly, a bespoke family recipe cutting board, a premium wood chopping board adorned with a treasured family recipe in beautiful handwritten script. Secondly, a custom-designed rolling pin, accompanied by a collection of gourmet ingredients including top-notch Madagascar vanilla beans, saffron, unusual spices, and unique sugars.

Lauren Goode: Thus, it didn't just offer suggestions, it also recommended places to purchase them. Initially, it didn't give direct links, but upon further inquiry, it did. I chose to buy the unique items, like Madagascar vanilla, saffron, and even gold flakes among others, thinking they'd be useful. The rest seemed a bit over-the-top.

Michael Calore: Indeed.

Zoë Schiffer: To some extent, yes, though I'm uncertain. However, the suggestions seemed quite beneficial to me. It's understandable why this particular advice stood out as the most useful, given that the stressful aspect of the process you mentioned initially isn't really about the tediousness of making a purchase. Instead, it's the struggle of coming up with unique gift ideas for others. So, the fact that these recommendations were very detailed and not something I would have immediately thought of myself, I find that quite intriguing.

Lauren Goode: Precisely. It's not a matter of simply hastening the purchase by 30 seconds with a Buy button. Rather, it's about being swamped with the brainstorming phase and seeking inspiration. Naturally, I did browse the web a bit more afterwards, but it effectively organized my options. It helped me decide on what to pursue.

Michael Calore: Additionally, that's the top vocal quality.

Zoë Schiffer mentioned, "I was about to comment on how impressive the voice quality was."

Lauren Goode: That's Cole for you. They've got these unique names, like Jasper, Cole. Perhaps Cole could help with sleep training your infant.

Zoë Schiffer: I'm aware.

Michael Calore expressed his dislike for the artificial insertion of hesitations such as "ums" and "ahs" in AI-generated voices, stating that they're unnecessary. However, he mentioned that Cole's voice quality was satisfactory

Lauren Goode: Absolutely. Shopping for my 16-year-old niece proved to be quite the humorous challenge, as I could vividly imagine the unimpressed looks she'd give me for choosing something completely out of touch. So, I put together a brief profile of her. She's in her junior year of high school, an avid reader with a beloved cat, and harbors aspirations of becoming a nurse. She's also busy preparing for her SATs. For gift ideas, I consulted Gemini, but the experience left much to be desired. Initially, it suggested a cat blanket, which puzzled me. I was uncertain if it meant a blanket adorned with cat images or a blanket intended for her cat – a known no-go area since cats typically prefer cardboard boxes to any cozy blanket you might provide.

Zoë Schiffer: It's like you're saying, "Gemini, has Gen Z ever unleashed their fury on you? I'm searching for fresher, more inventive concepts. Let's get moving."

Lauren Goode: Following that, the suggestion was to obtain an SAT guide for her.

Zoë Schiffer exclaimed, "Oh my God."

Lauren Goode expressed, "I was thinking, 'Wow…' I'd completely jeopardize my reputation as the cool aunt if I went ahead with that. They refer to me as the famous auntie, by the way, since they're aware I host a podcast and have a verified badge on certain social platforms. They find it incredibly cool. And I'm there thinking, 'Guys, that really doesn't hold any real value. It isn't.'"

Michael Calore: Right, so you should hand over eight dollars.

Lauren Goode: I ended up giving them quite a bit more than $8. It was a combination of cash and a gift card for both my niece and nephew, and they seemed really happy with it. It allows them the freedom to purchase whatever they desire, and they're at an age where the excitement of unwrapping presents isn't as appealing as it is for younger children. So, in terms of Gemini, that was a miss.

Michael Calore: From what you're saying, it seems like the experiences aimed at teenagers are lacking. Similarly, the ones that just throw a bunch of links at you aren't impressive. However, when it comes to generating new and unique ideas that wouldn't typically cross your mind, it actually proves to be quite effective.

Lauren Goode: Indeed, it's beneficial for generating ideas. However, it hasn't reached the stage where it operates independently, completing tasks overnight. That might be what we see in the future. Moreover, these chatbots seem to be in a phase of figuring out their exact purpose. Currently, Perplexity Pro's approach to shopping doesn't quite offer a compelling reason to do all one's shopping there, even though that's its ambition. It's essentially making its initial foray into the realm of retail. For them, the success metric seems to be if users spend a bit more time on Perplexity Pro compared to Amazon or Google, even if they eventually make their purchases elsewhere, they consider it a win.

Zoë Schiffer: Did you find the process to be more relaxed compared to the usual holiday shopping ordeal?

Lauren Goode: Honestly, no. I was behind on everything. I postponed several tasks, thinking, "I'll handle these next year." The special ingredient my friend was waiting for arrived post-holidays. Additionally, I attempted a trial with product returns, which we're all aware is a particularly tedious aspect of buying things. It's intriguing to imagine a future where, perhaps through interacting with chatbots or based on our previous purchases, we could simply say, "I need to make a return," and it would be taken care of for us. It would know the order details, manage the process, and maybe even print the return label on the spot. However, we're not there yet. Far from it. So, essentially, it's not proactive. It doesn't shop for you; it's more useful for generating ideas and doesn't assist with, for lack of a better term, the logistical follow-through. At this stage, it's essentially just a preliminary service.

Zoë Schiffer: Just.

Lauren Goode: Honestly, I'm still not a fan of shopping, but it's alright. Apologies, Cole.

Michael Calore: Alright, let's take a collective decision. On the whole, how satisfied have you been with utilizing AI to enhance your daily living? Zoe, your thoughts?

Zoë Schiffer: At this stage, I'd rate it a B minus. My hope is for it to excel, but currently, it seems merely satisfactory.

Michael Calore: Alright. Lauren?

Lauren Goode: My assessment is a firm B. The inclusion of ChatGPT significantly improves the rating, altering the standard.

Michael Calore: Good.

Lauren Goode: How about your thoughts, Mike?

Michael Calore: My rating would also be a B. I quipped about it earlier, but honestly, if you're adept at using the internet and searching online, then Gemini's insights aren't necessary for you. However, for those who are intrigued by the idea of interacting with technology and are novices in navigating this fresh terrain, it serves its purpose adequately. It didn't offer me anything particularly negative. The salad, too, gets a B from me.

Lauren Goode: Oh, my.

Michael Calore: That wraps up today's episode. Join us again next week for a discussion on the evolution of content moderation, its current state, and future directions. Thank you for tuning into Uncanny Valley. If you enjoyed our episode, please follow us and leave a rating on your preferred podcast platform. For any inquiries, feedback, or ideas for future episodes, feel free to reach out to us at uncannyvalley@wired.com. Today's episode was produced by Kyana Moghadam, with mixing by Amar Lal at Macro Sound. Our executive producer is Jordan Bell, and Chris Bannon serves as the head of global audio for Condé Nast.

Suggested For You…

Direct to your email: A selection of our top stories, curated daily just for you.

Live video streams and vehicle information are being exposed by license plate scanners

Cash Galore: Explore our exclusive edition right here

Top 17 Electric Vehicles Arriving in 202

Visualization: Monitoring all artificial intelligence copyright litigation in the United States

Additional Content from WIRED

Critiques and Manuals

© 2025 Condé Nast. All rights reserved. Purchases made via our website may result in a commission for WIRED, as part of our affiliate agreements with retail partners. Reproduction, distribution, transmission, caching, or any other form of usage of the site's content is strictly prohibited without the explicit consent of Condé Nast. Advertisement Choices

Choose a global website

The Sports Article You Opened Might Be Generated by AI

Before you dive into reading sports updates on the web, take a second look at the website address.

Take, for example, a title such as "Red Sox Encouraged to Skip Alex Bregman for a $427 Million Phenomenon" which might appear normal and seemingly originates from BBC Sports. However, upon a detailed examination, it turns out to be a counterfeit version named “BBCSportss,” and the content is plagiarized from Sports Illustrated. Additionally, within the same site, there are articles that haven't been directly copied from another author but rather seem to be a jumbled rehash of various sports bloggers' content, likely produced by artificial intelligence.

DoubleVerify, a company specializing in monitoring digital advertisements and media metrics, recently carried out a study on a group of more than 200 websites. These sites appeared to blend content generated by artificial intelligence with pieces of news articles lifted from real news sources. The study found that these websites frequently selected domain names and crafted their site designs to resemble those of well-known media entities such as ESPN, NBC, Fox, CBS, and the BBC. A significant number of these counterfeit sites present themselves as authentic sports news platforms.

Paige Graham, representing Sports Illustrated, stated, "The content under the name 'BBC Sportss' did not receive our approval and is actually a case of plagiarism."

"According to Dan Slivjanovski, the chief marketing officer at DoubleVerify, the rate at which fraudulent activities are increasing is alarming, with a year-over-year doubling or even tripling in occurrences. A significant portion of this fraud is attributed to bots that consume rather than produce content, artificially inflating website traffic to illicitly gain advertising dollars by simulating an increase in genuine viewer engagement. Additionally, DoubleVerify has noted a significant rise in the manipulation involving AI click farms. These are operations where networks of websites are filled with content generated by artificial intelligence, aiming to divert both legitimate readers and advertising revenue from authentic news sources."

DoubleVerify has identified a network dubbed "Synthetic Echo," which is known for replicating content from various sources, utilizing AI to create articles, or employing a combination of both strategies. Gilit Saporta, the head of the organization's fraud detection division, commented, "It's not even about spreading false information. It's essentially just incoherent content." According to Saporta, while this isn't the first instance of such a content scheme she has come across, what makes this particular case stand out is the apparent consistency in web design across multiple sites, indicating a clear connection between them.

WIRED requested Reality Defender, a company specializing in identifying deepfakes, to examine a selection of sites from the Synthetic Echo network. Upon reviewing the latest posts from "NBC Sportz," the firm concluded that although the material was originally created by human authors, it was plagiarized from authentic news platforms. The original authors' names were removed and replaced with "nbcsportz" as the source.

A website was found to have illegally republished content from the Detroit Free Press without permission. The Detroit Free Press has stated that the republished work was not approved or given a license by them. "The unauthorized content does not reflect our brand or the quality of our journalism," mentioned Lark-Marie Antón, a representative for the Detroit Free Press. The organization intends to pursue legal measures in response.

In its evaluation, Reality Defender also examined NBCSport.co.uk and BBCSportss.co.uk, finding that all the analyzed articles from these sites were probably produced by artificial intelligence, a conclusion that supports DoubleVerify's research.

NBC Sportz remained silent and did not reply to any inquiries for feedback. Both NBCSport.co.uk and BBCSportss.co.uk lack publicly available contact details, leaving WIRED without a means to reach out. These websites, along with a site mimicking CBS News suspected by DoubleVerify to be part of the Synthetic Echo network, are all registered under the domain registrar Namecheap.

For a long time, unauthorized individuals have tried to benefit from the success of established media platforms by redistributing their content illegally. However, the advent of AI technologies has enabled these practices to spread more rapidly than before. “The creation of substandard content isn’t a novel issue,” Saporta observes. “Yet, the modern technologies available today make it far simpler to duplicate and expand on a much larger scale.”

The quantity of websites producing low-quality content through artificial intelligence has seen a significant spike annually following the surge in generative AI tools' prominence in 2023. In February, not long after WIRED initially covered the emergence of content mills powered by AI, the media oversight organization NewsGuard had pinpointed 725 websites inundated with AI-generated articles. By January 2025, this figure had escalated to at least 1,150 such sites.

"The quantity has increased," states Shouvik Paul, the chief operations officer at Copyleaks, an AI detection firm. "Many of these operations are run from abroad and are quite dubious, so how can you possibly stay ahead?"

Adding to the perplexity for the audience, several leading news websites have dabbled in producing news pieces crafted by artificial intelligence. (For instance, Sports Illustrated reportedly featured content created by AI, which, according to its owning corporation, was sourced from an external provider.) Elsewhere, individuals keen on exploiting domain names have acquired web addresses of struggling media outlets, reviving them as production houses for AI-generated articles, often substituting their once reputable reporting with insubstantial, automated drivel.

A number of these platforms have started causing actual misunderstandings; in October, a digital marketing agency released a computer-generated notice about a Halloween procession in Dublin, Ireland. Despite there being no scheduled event, large crowds of party-goers arrived in anticipation of celebrations.

Paul from Copyleaks likened the tactic employed by certain websites, which adopt the look and feel of legitimate media outlets to distribute low-quality content, to a form of phishing. He noted that some of these websites go beyond mere imitation, engaging in actual phishing activities. For instance, DoubleVerify discovered a site within this network that mimicked a Nigerian Fox news outlet, bombarding visitors with dubious pop-up advertisements for software.

Despite the questionable appearance of the pop-up ads, the websites associated with this cluster seem to engage heavily in programmatic advertising. This type of advertising relies on automated, large-scale purchases instead of direct negotiations between specific websites and advertisers. These sites are often filled with banner ads managed by well-known programmatic ad servers such as Criteo and Sharethrough, both of which did not reply to inquiries for comments. According to a report by DoubleVerify, the creators behind Synthetic Echo strategically selected sports content as a primary focus due to its perception as being safer for brands compared to hard news.

While WIRED was observing various websites, automated advertisements from several leading firms were noticed, including technology leaders such as Asana and Oracle, online retail heavyweight Net-A-Porter, cosmetics powerhouse Sephora, and vacation destination Kalahari Resorts. When approached for a statement, none of these entities provided a response.

In an era characterized by dwindling confidence in the press and decreasing income for numerous news organizations, the emergence of these low-quality content factories represents a twofold problem. They contaminate the realm of information with inferior and plagiarized material, while also diverting advertising dollars away from genuine content creators.

Consider Exploring Further…

Direct to your email: Subscribe to Plaintext for Steven Levy's in-depth perspective on technology.

Perhaps it's the right moment to purge outdated conversation records.

Cash, Cash, Cash: Explore our exclusive edition here

The antics surrounding memecoins are only beginning.

Mysterious Depths: A Behind-the-Scenes Examination of Silicon Valley's Impact

Additional Content from WIRED

Evaluations and Manuals

© 2025 Condé Nast. All rights reserved. A share of revenue from products bought via our website may go to WIRED, as a result of our Affiliate Agreements with retail partners. Content from this site cannot be copied, shared, sent, stored, or utilized in any form without explicit prior consent from Condé Nast. Choices in Advertising

Choose a global website

AI Assistants Have Arrived. What Limits Should We Set on Their Capabilities?

Is it wise for me to employ a personal AI assistant to aid in managing my everyday activities?

—In Search of Help

Generally speaking, I believe that depending too much on any form of automation in our everyday activities can be risky if overdone, and can even feel isolating if used sparingly, particularly when it comes to interactions with others. Having an AI assistant that manages my schedule and compiles links for additional research? Wonderful. But an AI assistant that sends automated updates to my parents on a weekly basis? Terrible.

The most compelling reason to limit the integration of additional generative AI applications into your everyday activities is their ongoing environmental toll during both the training phase and the creation of outputs. Bearing this in consideration, I explored the archives of WIRED, from the early, chaotic days of what we now know as the internet, to provide a deeper historical background for your inquiry. After a thorough search, I returned with the conclusion that it's very probable you're already engaging with AI tools on a daily basis.

The concept of AI agents, or the more daunting term "agentic AI," has become the latest trend among tech executives eager to promote their latest investments. However, the notion of a digital helper designed to carry out software-related tasks is hardly new. The discussions that took place in the 1990s about "software agents" closely resemble today's talks in Silicon Valley. Here, tech moguls are forecasting a surge of generative AI agents geared towards handling our internet tasks for us.

In a 1995 WIRED interview, MIT professor Pattie Maes expressed concerns over accountability regarding the behavior of agents. She highlighted potential issues, such as agents occupying excessive machine time or making unwanted purchases. Despite these challenges, Maes firmly believes that agents will become indispensable, stirring up numerous intriguing questions along the way.

In the early days of January, I reached out to Maes to get her current views on AI agents and how they've evolved. Her enthusiasm for the possibilities of personal automation remains high, yet she firmly believes that engineers, whom she describes as "extremely naive," are neglecting the intricate aspects of how humans and computers interact. She warns that such negligence might lead to another period of stagnation in AI development, reminiscent of the AI winter.

"She points out that the current construction of these systems prioritizes technical and engineering efficiencies," she explains. "However, they fall short when it comes to considerations of human-centric design." Her emphasis is on the susceptibility of AI agents to deception or to adopting prejudiced views, even as the fundamental models see enhancements. This overconfidence in the systems causes users to place undue trust in the outputs of AI technologies, often when it's unwarranted.

To gain a clearer insight into other possible challenges for personal AI assistants, we can divide the broad concept into two separate groups: the ones that provide information to you and the ones that act on your behalf.

Feeding agents utilize algorithms that analyze your preferences and behaviors to sift through vast amounts of data to pinpoint what matters to you. This concept probably rings a bell. Think of it as the technology behind social media platforms that curate your feed with posts you'll likely enjoy or the relentless advertising algorithms that keep showing you the same mushroom gummy ads on Instagram. These can be seen as your own AI agents working to personalize your online experience. For instance, in an interview from the 1990s, Maes talked about a news-collecting agent designed to filter and present articles of interest to her, much like how my Google News page seems to know exactly what I like to read.

The problem with overly depending on AI assistants is the gradual onset of comfort that leads to indifference. The more I allow algorithms to tailor the content and recommendations for me, the more likely I am to fall into a routine of predictable choices, encountering fewer unexpected delights. Should I rely on an AI to organize my weekly grocery shopping and utilize a delivery service for my purchases, initially, the efficiency might appeal to me. However, over time, the repetitive nature of this convenience could discourage me from trying out new recipes. Thus, life slowly turns into a journey of increasing dullness.

Regarding AI entities acting as your proxies, they are fundamentally just complex software routines. Their purpose is to observe and emulate your actions, thereby enabling them to represent you in various online settings. Imagine a virtual counterpart capable of participating in video conferences, crafting gratitude messages for your wedding attendees, and making dinner reservations on your behalf. This category also encompasses advanced AI applications designed to mimic celebrities and automatically engage with their followers.

Steer clear of utilizing or engaging with these AI duplicates, regardless of the situation. The virtual realm has surpassed the tangible one in significance, with much of our existence now filtered through digital interfaces. Although technology might improve to the point where deploying an AI to video call my grandmother or allowing it to participate in a group chat with my buddies doesn't seem absurd, we ought to reserve automation for monotonous activities. It's crucial to be present and authentic in our digital communications, as it's becoming the predominant way we connect.

Participate in Discussions

Become a part of the WIRED family by contributing

Something to Check Out…

Direct to your email: A selection of our top stories, carefully chosen for you daily.

Live video streams and automobile information are being exposed by license plate scanners.

Cash Galore: Explore our exclusive edition now

The Top 17 Electric Vehicles Arriving in

Tracking Visualization: Monitoring all AI-related copyright legal cases in the US

Additional Content from WIRED

Insights and Tutorials

© 2025 Condé Nast. All rights reserved. Purchases made via our website may result in a commission for WIRED, as part of our affiliate agreements with retail partners. Content from this site cannot be copied, shared, broadcast, stored, or used in any form without explicit consent from Condé Nast. Ad Choices

Choose a global website