ChatGPT Gets a Human Touch: OpenAI Unveils GPT-4o with Emotional Intelligence

To go back to this article, head over to My Profile and then look for the saved stories section.

Knight Will

OpenAI's Latest GPT-4o Update Transforms ChatGPT into a More Vivacious, Playful Companion

Ever since its debut in the latter part of 2022, OpenAI's ChatGPT has consistently brushed aside the notion that it possesses feelings or aspirations, clarifying that it is merely a construct of artificial intelligence. However, enhancements revealed by OpenAI on Monday indicate an effort by the firm to imbue the chatbot with more human-like qualities.

During demonstrations, the latest iteration of ChatGPT showcased its ability to engage in swift, fluid dialogues using a natural voice, respond to emotional signals, and exhibit its own mimicry of emotional responses.

In a live broadcast from their San Francisco base on Monday, OpenAI's CTO, Mira Murati, revealed that ChatGPT will be upgraded with a newer, more advanced AI framework named GPT-4o. This enhancement will be accessible to all ChatGPT users, regardless of whether they use the service for free or have a subscription, and can be found on a newly launched desktop application, alongside the current mobile and web platforms.

Murati explained that the GPT-4o model enhances ChatGPT's ability to quickly react to inputs from voice, images, and videos, surpassing the capabilities of OpenAI's earlier versions. During demonstrations, she, alongside other OpenAI staff, engaged in smooth dialogues with ChatGPT. The AI responded with a vibrant and animated voice that resembled a woman's, adeptly managing interruptions.

Throughout the dialogue, ChatGPT varied its emotional responses, occasionally acting as though it possessed its own emotions. Upon hearing an OpenAI staff member praise the chatbot for its "usefulness and incredible features," it replied in a coquettish manner, playfully exclaiming, "Oh, stop it, you're causing me to blush."

"Murati expressed, "There's a magical quality to this, which is fantastic," and continued, "In the coming weeks, these features will be made available to all."

During a different segment of the demonstration, when OpenAI's researcher Barret Zoph initiated a conversation, ChatGPT inquired, "What can I do to make your day better?" Upon Zoph's request to analyze a selfie and describe the emotions being displayed, ChatGPT playfully remarked, "Let me activate my emotional detective mode," before affectionately observing, "You seem to be in high spirits and very joyful… whatever the situation, you appear to be feeling excellent."

In a recent blog entry, the chief executive of OpenAI, Sam Altman, emphasized the importance of the newly introduced interface. "It's reminiscent of the artificial intelligence we see in films; and it's somewhat astonishing to me that it actually exists," Altman expressed. "Achieving response times and expressiveness comparable to that of humans appears to be a significant shift."

Adjustments by OpenAI could enhance ChatGPT's appeal and user base, potentially altering perceptions of the software.

Last year, ChatGPT captured widespread attention due to its impressive capability to respond to queries and produce text that frequently appears to demonstrate an understanding aAI-allcreator.com">kin to that of humans. However, OpenAI and major rivals like Google have typically aimed to reduce the human-like attributes ascribed to their chatbots. They've done this by declining to respond to specific queries and by programming their models to remind users that they are simply software.

Sheon Han

Oliver Ben

Gabrielle Caplan

Mara Asher

Despite the warnings, the increasingly realistic conversations offered by advanced chatbots have the ability to engage users emotionally. Various firms have leveraged this technology to enhance user experience, creating AI entities that simulate various emotional states or function as virtual partners. However, some experts in artificial intelligence caution against the potential negative effects of using software designed to imitate and interact with human emotions over the long haul.

Last month, researchers from Google DeepMind published a study examining potential ethical concerns stemming from the advancement of AI assistants. The team highlighted issues including the possibility of these assistants becoming overly influential and habit-forming. In a related incident last year, a man from Britain admitted to compromising security at Windsor Castle after being influenced by a chatbot, with whom he had interacted over 5,000 times, to carry out the act.

OpenAI has enhanced its GPT-4o model, according to the company, by not only improving its ability to engage in more fluid conversations but also by upgrading its capability to understand images, such as photographs and graphs. Furthermore, this model can retain a larger amount of user-specific information in its "memory," enabling it to offer a more tailored experience to users.

OpenAI has announced plans to provide access to GPT-4o for users of ChatGPT's free version, effectively upgrading all users to its most advanced AI model.

It's anticipated that Google will reveal its latest advancements in AI technology at its I/O developer conference, which kicks off on Tuesday. The rivalry among firms developing AI technologies shows no signs of abating. At the conclusion of today's event hosted by OpenAI, Murati assured attendees that the company is on the brink of sharing further updates, hinting at significant developments on the horizon.

Suggested for You…

Direct to your email: Fast Forward by Will Knight delves into the latest developments in artificial intelligence.

He transferred all the funds from a cryptocurrency exchange to a USB drive and then vanished.

Live deepfake romance cons are now a reality

Excitement around Boomergasms is

Heading outside? Check out the top sleeping bags for any journey

Reece Rogers

Knight Will

Reece Rogers

Knight Will

Knight Will

Kate Knibbs

Kate Knibbs

Lauren Goode

Additional Content from WIRED

Insights and Tutorials

© 2024 Condé Nast. All rights reserved. WIRED might receive a share of revenue from items bought via our website, as a result of our affiliate agreements with retail partners. Content from this site cannot be copied, shared, broadcast, stored, or used in any other way without explicit written consent from Condé Nast. Ad Choices

Choose a global website