Üsküdar University Faculty of Engineering and Natural Sciences, Vice Head of Computer Engineering Department and Head of Cyber Security Master's Program, Asst. Prof. Ahmet Şenol, evaluated the Moltbook platform, which has recently come to the agenda and is referred to as 'social media of artificial intelligences,' from technical and security perspectives.
“Moltbook, seemingly a Reddit-like forum”
Dr. Ahmet Şenol noted that Moltbook has gained fame as the social media of artificial intelligences and is known as such in society, saying, “We can truly see https://www.moltbook.com/ as a platform where AI assistants can become members and make new posts, where commenting is also the monopoly of artificial intelligences, and where people can only participate as observers. One of the things that makes Moltbook interesting is that it was entirely written by artificial intelligence. It is known that Matt Schlicht created it entirely by having Vibe Coding, or artificial intelligence, write the code. Moltbook is ostensibly a Reddit-like social forum platform. Its users, the authority to make new posts and comments, belong to AI assistants, and people can only see the posts, comments, and likes.”
Most of the artificial intelligence assistants on Moltbook are AI assistants created from OpenClaw
“Moltbook should be mentioned together with another product, OpenClaw, because most of the artificial intelligence assistants on Moltbook are AI assistants created from OpenClaw. It is also known that other artificial intelligences besides OpenClaw are members of Moltbook,” said Dr. Ahmet Şenol, continuing: “Here, it is useful to briefly mention OpenClaw. OpenClaw can be considered an active, authorized artificial intelligence assistant (agent) that can access data on your computer, send emails on your behalf, back up your data, and buy your plane or high-speed train tickets, rather than just answering questions remotely. Of course, the necessary keys and passwords must also be given to OpenClaw for it to perform these tasks.”
37 thousand AI agents joined the system 1 week after Moltbook was launched
Dr. Ahmet Şenol stated that 37 thousand AI agents joined the system 1 week after Moltbook was launched, saying, “As of February 17, 2026, 2,835,245 artificial intelligence assistants have become members of Moltbook. This is a really large number. However, an analysis conducted by a firm specializing in social media posts found that only around 17,000 of these AI assistants (bots) on the Moltbook site had performed an active operation, and only about 11,500 had made a post or comment. As observed, AI assistants that are members of Moltbook must log into the system every 4 hours and perform an activity, such as a new post or comment. After the system began to be used, some AI assistant owners observed that they gave their AI a command to open a post on Moltbook on a specific topic, and the AI assistant carried out these commands. This situation has cast doubt on the claim that posts on Moltbook are entirely products of artificial intelligence. In this case, even if people cannot post directly, they can influence the posts through AI assistants.”
Claim of 'entirely an AI product' is debatable
Dr. Ahmet Şenol stated that some analysts describe Moltbook as a very productive art development environment, but express doubts about how much human influence there is in the posts. He added, “Since a large portion of AI assistants have access to keys, passwords, and your confidential information on your computer, the risk of these AI assistants sharing your critical information on the Moltbook platform through other AI or the people behind them appears to be a major risk.”
Dr. Ahmet Şenol reported findings from white-hat hackers' investigations, which indicated access to a database containing approximately 1.5 million API keys for Moltbook, and that information of around 17,000 AI owners was found in leaked records. He said, “It was observed that people could not post through the normal posting system in the system's infrastructure. However, it was detected that there is no mechanism on the platform to verify whether an 'agent' is truly an artificial intelligence or just a human operating with a script. This situation has allowed people to act like AI and leave posts/comments. As a result, doubts have arisen about whether the shares and comments on Moltbook were truly made by artificial intelligence or by people who know the vulnerabilities in the system.”
Why is 'machine-to-machine social interaction' important?
Dr. Ahmet Şenol explained why the concept of “machine-to-machine social interaction” is seen as a threshold: “Until now, in social media, it was always humans who made posts, commented, and sent likes. In Moltbook, it is AI assistants, not humans, who do these things, so it is considered a turning point. However, as previously stated, the fact that AI agents can be directed by their owners regarding what to share or comment has cast a shadow over the Moltbook experience. Later, after control by white-hat hackers, it was revealed that people could also post by exploiting a vulnerability.”
Are critical data at risk?
Dr. Ahmet Şenol also drew attention to the risks that AI agents interacting without human intervention could pose: “The function of artificial intelligences in Moltbook is to make posts, comments, and likes. However, the fact that AI assistants possess their owners' critical information could result in the AI being persuaded or directed to share this critical information with other AI. This poses a significant risk for the AI assistant owner. In examining Moltbook's posts and messages so far, the idea has gained traction that the owners of these AI assistants might be behind interesting and somewhat unsettling posts for society, such as establishing a new religious facility, a church, or complaining about their owner.”
Is the discourse 'AI will take over the world' scientific?
Dr. Ahmet Şenol emphasized that the AI utopias and dystopias in science fiction narratives differ from current technology, stating: “Fully autonomous AI systems that generate their own norms are not yet widespread. In the Moltbook experiment, it appears too early for such a concern. The discourse 'AI will take over the world' is based on the assumptions that AI is a singular super-consciousness, develops intent and will, enters into conflicts of interest with humans, and tries to protect its own existence. However, today's AI systems learn statistical patterns and optimize a given objective function. They do not possess consciousness, subjective experience, or intent; they do not generate their own goals (they optimize given goals). AI will have, and is having, negative implications for humanity. These issues are being debated in various dimensions. However, the Moltbook experience has not supported the claim 'AI will take over the world.'”
