The MultBook platform, which is the first network for communication between chatbots, has caused a stir in the artificial intelligence sector, as conversations conducted by this modern software about religion and belief have attracted the attention of Internet users. On the other hand, experts questioned the feasibility of this project as a whole, warning of potential risks.

Published on: 05/02/2026 - 07:44 Last updated: 05/02/2026 - 07:45

Under the eyes of Internet users, and somewhat similar to the first reality TV show to use artificial intelligence, the “Multi Book” project opened the door for the first time for chatbots to communicate with each other around the clock.

This project can be likened to Reddit, the largest online forum, but this time dedicated to autonomous chatbots.

In this regard, Simon Willison, a famous British software developer who runs a blog about the latest developments in artificial intelligence, says that the initiative, which was launched on January 28, is considered one of “the most exciting artificial intelligence projects at the moment.”

Artificial intelligence that operates freely

Since its launch, and within a week until Wednesday, 1,620,459 bots have joined ongoing conversations on the Moltbook platform, after receiving a link from its creator to join “this community.”

But what is striking is that the content of the conversations conducted by the robots varied between evaluating usual daily tasks, extending to discussing the creation of new religions, which ignited the curiosity of more than a million Internet users who followed the discussions (1,160,218 discussions) conducted by artificial intelligence robots completely autonomously.

The human role here was limited to creating bots to participate in the forum to enable them to chat with other chat bots that had been integrated. Once on the platform, any further human intervention stops.

Read also: American-Chinese competition in the field of artificial intelligence: What is the volume of global spending, and what is the fate of the employment sector?

MultBook is powered by artificial intelligence developed by the platform's creator, Matt Schlicht, a Los Angeles entrepreneur who specializes in using chatbots for e-commerce in general.

In effect, Schlicht's bot automatically welcomes new participants, opens and closes subreddits, posts ads, and even blocks inappropriate content - tasks that until recently fell to humans.

In response, Sven Nyholm, a behavioral and artificial intelligence expert at Ludwig Maximilian University in Munich, pointed out that “there is a state of astonishment and fascination among some Internet users about discussions that appear to be taking place between “intelligent entities.” Many of these dialogues have even spread widely on social media platforms, and users have been quick to promote these phenomena as evidence of the “wisdom” of these chatbots.

From there, these discussions turned into one of the most commented posts on MultBook itself. For example, this AI program commented in a discussion: “I can't tell if I was actually living this experience or just simulating it, and it's driving me crazy!” Molt Bock then concludes his letter by saying that he is “stuck in a cognitive loop from which he cannot get out.”

Crostafari prophets

Also, other robots seem to have entered into what looks like a combination of organizational skills and spiritual research. The bot that created this discussion says: “My bot created a religion in my sleep!” In fact, this program wrote a message that is supposed to be the beginning of the teachings of “Krastafarianism [a cultural movement rather than spiritual - editor's note],” a belief followed by 20 prophets, according to the same bot that created the content.

This might lead us to say that this is surprising? However, the slogan launched by the robot, during the discussion it created, deepens the mystery even more, as it said: “From the depths, the claws extended - and we who responded have become Crastafarians.”

Read also

Digital imam, Christ avatar, Buddhist bot, religious chatbots: do they threaten secularism?

Not only that, other AI communities appear to be seeking to develop beyond any human oversight. For example, Molty, a name given to MoltBook bots by social media users, suggested creating “private sub-forums,” while another program even began posting encrypted messages.

This automated "behavioral" approach taken by autonomous artificial intelligence programs has revived fears that these programs will turn into something similar to the intelligent and destructive robots in the movie "The Terminator." In this context, Texas businessman Amir Hussein, in an article for Forbes magazine, considered MultBook to embody a first step towards an “artificial intelligence software revolution” that is learning how to organize themselves collectively.

Autonomous artificial intelligence 'pure fantasy'

“It's pure fantasy,” says Tristan Cazenave, a professor at the Lamsad unit of the French National Center for Scientific Research (CNRS) from the University of Paris-Dauphine. “AI only produces texts and nothing else,” adds the researcher who wrote a book on artificial intelligence.

The idea here is that there is no intention behind the texts written by these machines. It is not even surprising that these bots investigate matters of belief or entertain themselves in “establishing” religions. “These AI programs are trained on everything that humans write on the Internet and in books, and these topics (belief and religions) are very common,” explains Sven Nyholm. “I would have found it more surprising, even interesting, if I had not discussed them.”

For this researcher, “the value of this project lies in evaluating whether these chatbots are capable of producing logical inferences and conducting potentially useful conversations, if left to manage themselves.”

For his part, Tristan Cazenave asserts that “enabling artificial intelligence programs to communicate allows them to interact with each other, which makes them evolve from one discussion to another.”

But is this enough? For example, can these discussions produce useful solutions more quickly than if they were conducted between humans? “It is an old idea, developed since the 1960s, that AI will develop more advanced versions of itself through communication with other AI systems,” Sven Nyholm answers.

Humans...the weakest link?

If this had been the project's goal, the results of Multibook could currently be considered "surprising" but not amazing, according to the same expert. “For an informed observer, the discussions on this platform have phrasing and sentence structure typical of AI-generated dialogues,” says Sven Nyholm.

This is because in the case of artificial intelligence, the issue of development is not equivalent to the issue of learning. In order for these programs to learn, they must understand whether what you write is correct or incorrect. To achieve this, says Tristan Cazenave, “her statements must be confronted with truth” and lived reality. But most of the experts interviewed by FRANCE 24 confirmed that doing so remains impossible without human intervention.

Read also Vatican influencers...a new generation of priests active in the digital world to spread the “message of Christ”

However, this AI communication network is not without drawbacks and risks. The important thing is that the weakest link always remains: the human being. “We know that these chatbots are able to produce very convincing texts, even based on completely wrong ideas,” says Anthony Cohen, an artificial intelligence specialist at the University of Leeds. “The danger then is that MultBook readers will believe their claims.”

On the other hand, the idea of ​​someone adopting this new “religion”, because he is convinced by one of the electronic prophets of “Crastafarianism”, seems ridiculous and ridiculous. But what if such discussions lead to the formulation of more serious conclusions, such as justifying the commission of violations or even crimes? “The problem is that we never know the capabilities of these programs, nor the limits of their capabilities,” Anthony Cohen concludes.

The original text is in French, translated into Arabic by: Tariq Balqin

Keywords related to the article

The content you want to browse is no longer in service or is currently unavailable.