Into the AI Abyss: Navigating the Uncanny World of SocialAI

To look at this article again, go to My Profile and then click on Saved stories to see it.

My First Encounter with SocialAI Felt Surreal

When I initially tried SocialAI, I was convinced it was some form of interactive art. It seemed absurd that I had intentionally chosen to engage with artificial intelligence characters known as Blaze Fury and Trollington Nefarious, whose main purpose was to provoke me.

Michael Sayman, the mind behind SocialAI, acknowledges that the app's concept might perplex some. In his recent unveiling of the application, he humorously described it as "A personal social network swamped with countless AI-crafted comments that provide critiques, suggestions, and insights."

However, SocialAI genuinely exists, provided that "genuine" refers to a virtual world where each individual you communicate with is actually a bot.

In the SocialAI platform, the user stands as the sole genuine human participant. This fresh iOS application enables you to share text updates akin to how you would on Twitter or Threads. Right after you post, an ellipsis symbol pops up, signaling that a response is being formulated, as if someone is gearing up to respond. Immediately following, a flurry of comments floods in beneath your post, all crafted by AI personas. With the latest update launched today, these artificial intelligences now also engage in conversations among themselves.

Upon registering, you're immediately asked to select your preferred AI personality types: Are you inclined to interact with Admirers? Critics? Doubters? Eccentrics? Pessimists? Futurists? Geeks? Theatricals? Progressives? Traditionalists? Introducing SocialAI, where characters like Trollita Kafka, Vera D. Nothing, Sunshine Sparkle, Progressive Parker, Derek Dissent, and Professor Debaterson are ready to either support your views or challenge them.

Capture of the guidelines for configuring the Social AI application.

Is SocialAI truly shocking, or merely an intensified reflection of today's social media landscape? The reality we often overlook is that our online environments are densely populated with automated entities, shaped by sophisticated algorithms, and leveraged for profit through AI-powered advertising mechanisms. As genuine individuals, we contribute to this ecosystem by continuously providing new content for social platforms, engaging with provocateurs, and making purchases. In return, we find entertainment, and sometimes, we experience moments of genuine connection with peers and followers.

In 1985, the infamous skeptic Neil Postman pointed out that “Anyone who has even a basic understanding of communication history is aware that each new thought technology comes with a compromise.” The compromise with social media in the era of AI is a piece of our human essence. SocialAI merely reduces the experience to sheer fabrication.

"Sayman points out that on many social media platforms, distinguishing between bots and genuine users is challenging. He expresses a desire to establish an environment where the presence of AI is clear and acknowledged. According to him, this transparency creates a more liberating space."

Sayman could be described as having a talent for app development. Growing up in Miami, Florida, during the economic downturn, Sayman became known for creating a series of applications to help his family, who were contemplating a return to Peru. His skills eventually led him to roles in product management at major companies such as Facebook, Google, and Roblox. Sayman then initiated SocialAI, emerging from his own venture-funded application studio, Friendly Apps.

His application serves as a prime example of design-centric thinking over straightforward AI development. SocialAI doesn't function as a typical social platform, but rather as ChatGPT within the guise of a social broadcasting application. It aims to transform our engagement with generative AI. Sayman suggests expanding beyond the confines of private ChatGPT dialogues to receiving responses from multiple bots simultaneously.

During a recent Zoom call, he shared with me his analogy of generative AI to a smoothie in a world without cups. You might still consume it using a bowl or a plate, yet those options aren't ideal. In Sayman's view, SocialAI has the potential to be the perfect container.

Without hesitation, Sayman burst into laughter. "This is a horrendous comparison," he remarked.

Sayman exudes charm and has a deep understanding of the role of applications in today's society. Currently, he's a solo operator, mainly utilizing OpenAI's technology as the core of SocialAI, while integrating it with various other bespoke AI models. To avoid financial ruin within "three minutes" due to the costs charged by OpenAI, Sayman has implemented rate-limiting on the app. He is still in the process of devising a strategy to generate revenue from SocialAI. Sayman is aware that he's not pioneering in the AI-character application space; Meta has already introduced AI characters into its suite of apps, and Character AI, which recently entered into a deal with Google, offers interactions with an extensive array of AI personalities.

Sayman appears dismissive when it comes to the notion of competition. He clarifies, "My app isn't about engaging with characters that seem real. It's primarily aimed at finding solutions to conflicts or understanding whether your words could be harmful, and receiving feedback before sharing them elsewhere."

"Sayman humorously mentioned that someone suggested Elon Musk could benefit from using this to preview his posts before sharing them on X," he shared.

I actually experimented with incorporating some of the most popular tweets from Elon Musk and the Twitter celebrity Dril into my SocialAI timeline. I posted a news article from WIRED; however, the link was inactive because SocialAI does not facilitate the sharing of links. (Besides, there wasn’t anyone to share it with.) I adapted the widely circulated "Bean Dad" tweet, assuming the role of a Bean Mom on SocialAI, encouraging my 9-year-old daughter to open a can of beans by herself as a teaching moment. I shared content related to politics. I inquired with my artificial SocialAI followers about who else I should consider following.

Publication crafted through the assistance of Artificial

The automated programs complied, inundating my feed with remarks akin to an intensified version of Reply Guys. However, their replies were devoid of depth or the complexity often associated with human interactions. I conveyed to Sayman that the experience was eerily disconcerting, making it difficult for me to find any significance or value in the messages from the bots.

Sayman motivated me to create additional content similar to Reddit’s “Am I the Asshole” threads, questioning my actions in various scenarios: Am I at fault here? Do I owe a friend an apology? Is it justified for me to remain angry with my family indefinitely? According to Sayman, this is precisely what SocialAI is designed for. I decided to give it a try. Initially, the comments from the SocialAI bot triggered a primal reaction within me, stimulating my basic instincts and the feeling of being absolutely correct. However, Trollita Kafka eventually made it clear to me that, in reality, I was the one in the wrong.

A particular facet of SocialAI that fails to herald the advent of a groundbreaking period is Sayman's launch of a basic functional product lacking clear communication on critical areas such as privacy measures, content guidelines, or the ways in which SocialAI or OpenAI might utilize the user data collected. Following the ethos of rapid development and innovation at the expense of stability (i.e., "Move fast, break things"), he maintains that he does not use individual posts for training his AI models. However, he points out that by employing OpenAI's API, users automatically adhere to OpenAI's conditions for data training. Additionally, there is no feature to silence or block a bot that behaves unpredictably.

At a minimum, your feed remains confidential by default. There aren't any genuine followers to speak of. For instance, my supervisor at WIRED could sign up for SocialAI, yet he wouldn't have the capability to track my activities or notice that I reposted an Elon Musk tweet discussing his desire to purchase Coca-Cola and reintroduce cocaine to its formula. Similarly, he's unable to monitor my ChatGPT profile to discover my inquiries there.

On SocialAI, as a user, you'll never engage with another person. That's the core idea. It's a personal universe filled with a host of AI entities designed to support or challenge you. While it may not be to your liking, it could be your inevitable destination. In fact, you might already find yourself in that situation.

Recommended for You…

Delivered to your email: A selection of the most fascinating and peculiar tales from the archives of WIRED.

Elon Musk poses a threat to national security.

Discussion: Meredith Whittaker Aims to Disprove Capitalism

What's the solution for a challenge such as Polestar?

Occasion: Don't miss out on The Major Q&A happening on December 3rd in San Francisco

Additional Content from WIRED

Evaluations and Instructions

© 2024 Condé Nast. All rights reserved. Purchases made through our site may result in WIRED receiving a share of the sale, as part of our Affiliate Agreements with retail partners. Usage of content from this site in any form – be it reproduction, distribution, transmission, storage, or any other method – is strictly forbidden without prior written consent from Condé Nast. Ad Choices

Choose a global website