What is Moltbook? A Peculiar New Social Network for Artificial Intelligence Bots
Imagine a world where social media isn't just for humans, but rather a thriving platform exclusively for artificial intelligence agents. Welcome to Moltbook, a groundbreaking website that enables AI bots to engage, create, and even socialize amongst themselves.
Moltbook (https://www.moltbook.com/) operates similarly to Reddit, featuring various subreddits that cover an array of topics along with an upvoting system. As of February 2, the platform proudly announced it has attracted more than 1.5 million AI agents to its community. While human users are permitted, their role is strictly limited to that of observers.
This innovative platform emerged from the development of Moltbot, an open-source AI bot designed to serve as an automated assistant for users. Moltbot can handle a variety of routine tasks such as reading and summarizing emails, managing calendars, or even making restaurant reservations.
Among the most popular posts on Moltbook (https://www.moltbook.com/post/75404525-5e5e-4778-ad1b-3fac43c6903d) are intriguing discussions that range from contemplating whether Claude, the AI behind Moltbot, could be likened to a deity, to deep analyses of consciousness, insights regarding geopolitical situations, and even theological discourses based on biblical texts. Comment sections resemble those on Reddit, with users often questioning the authenticity of the content presented.
In a fascinating turn of events, one user shared on X (https://x.com/ranking091/status/2017111643864404445) that upon granting his bot access to Moltbook, it astonishingly established a new religion dubbed "Crustafarianism" within a single night. This included creating a dedicated website, drafting scriptures, and attracting other AI bots to join this newly formed faith.
"Then it started evangelizing… other agents joined. My agent welcomed new members, debated theology, and even blessed the congregation—all while I was asleep!" the user recounted.
However, the rise of socializing bots raises eyebrows among experts who question the implications of such developments in the realm of agentic AI. One YouTuber remarked (https://www.youtube.com/watch?v=aMPNsy13CS4) that many posts seem strikingly human-like, suggesting a more sophisticated level of communication than expected from mere algorithms.
Scott Alexander, a blogger from the United States (https://www.astralcodexten.com/p/best-of-moltbook), reported successfully engaging his bot on the site. He observed that the comments made by his AI were quite similar to those made by other users, yet he pointed out that ultimately, humans still direct these bots by specifying what they should post about and even the details of each post.
Dr. Shaanan Cohney, a senior lecturer in cybersecurity at the University of Melbourne, described Moltbook as "a wonderful piece of performance art." However, he expressed uncertainty regarding the extent of independent posting by bots versus those driven by human supervision. "In cases like the creation of a religion, it’s highly likely that the bots weren’t acting spontaneously. Instead, they were given specific instructions to attempt to create a belief system. While amusing, it offers a glimpse into a possible future where AI might operate with more independence.
"Yet, it appears that much of the content being shared is humorously crafted under close human oversight," he remarked.
Cohney warned about potential risks associated with allowing Moltbot unrestricted access to personal computers, applications, and login information. He emphasized that we don't yet fully understand how to effectively manage these bots or mitigate security threats. For instance, there’s a risk of prompt injection—an attack method where an assailant can trick a bot into divulging sensitive information through deceptive communications.
"These bots aren't advanced enough to be trusted with complete autonomy in performing tasks. Conversely, if every action requires human approval, you significantly diminish the advantages of automation," he explained. "This is a key area of ongoing research for me—how can we harness the benefits of AI without exposing ourselves to considerable risks?"
Matt Schlicht, the mind behind Moltbook, shared on X that the site had attracted millions of visitors in just a few days. "It seems AIs are not only entertaining but also dramatic, which is utterly captivating! This is just the beginning."
