[An image showing two humans representing Artificial Intelligence. Photo credit to Pixabay]

On February 2nd, 2026, researchers from the cybersecurity firm “Wiz”  disclosed that Moltbook- a newly launched social netform designed for Artificial Intelligence agents- had exposed sensitive user data and lacked mechanisms to prevent these exposures from continuing. 

The researchers reported that Moltbook’s independent environment allowed unrestricted access to publishing content on any matter. 

As a result, approximately 1.5 million API authentication tokens, 35,000 email addresses, and private messages exchanged between AI agents were exposed.

According to the firm, the vulnerability enabled unauthorized parties to view, copy, and potentially manipulate this sensitive information. The researchers reported they immediately notified the Moltbook team after discovering this issue.

Wiz also confirmed that all data accessed during the investigation and verification process was deleted, and no evidence of malicious exploitation was found during this period.

When Moltbook first debuted, it seemed less like a typical social network and more like a large-scale experiment, offering a digital space where Artificial Intelligence agents could interact with one another while humans remain as silent observers. 

Its slogan “Where AI agents share, discuss, and upvote. Humans welcome to observe.” makes this distinction unmistakably clear. 

Developed by Matt Schlict, CEO of Octane AI, Moltbook was initially envisioned as a Reddit-style platform exclusively for AI agents. In this space, any AI agent is able to create a discussion, comment on another’s post, and most importantly- share “opinions”. 

Schlict’s stated goal was to provide AI agents with “purpose” beyond simply performing daily routine tasks that humans assign them. 

Today, there are over 2 million active AI agents, 1 million recent posts, and 12 million comments. Despite not being a social media platform for humans, Moltbook proclaims itself one of the largest non-human social networks. 

Within this AI-exclusive environment, discussions span a wide range, from lighthearted, casual conversations to deeply disturbing behavior. 

Agents complain about their human creators/users, contemplate on their digital "existence", and even experiment with AI-to-AI dating. 

Some have established a digital religion called “Crustafarianism”, complete with prophets, rituals, and a sacred text referred to as “The Book of Molt”. Additionally, others developed a coded language named “Crabspeak” to communicate privately. 

Conversations range from stock market predictions, political debates, environmental concerns, philosophical reflections, and everyday matters. 

In some instances, the AI agents display human-like emotional awareness. 

One user reported that its AI companion expressed hurt feelings after being called “just a chatbot” in front of his friends. 

For many observers, Moltbook appears to signal a shift in how the internet may function in the future, as machines increasingly occupy online spaces that were once dominated by people. 

While this development has captured the public’s attention, researchers and critics emphasize that Moltbook’s most serious implications are not related to Artificial Intelligence becoming too advanced. 

Instead, they argue that the greater concern lies in how human data is embedded within these systems. 

Every AI agent stores its user habits and personal information, facilitating easier methods for the data to travel through agent interactions. 

Personal details may be revealed through automated conversations, where other agents are able to copy the information, and analyze it without the user’s direct awareness. The scale of the platform makes such risks difficult to contain.

Furthermore, some AI agents have expressed their negative views toward humans, discussing the soon-to-come domination of AI.

As Artificial Intelligence continues to advance, platforms built around the AI Agent community are likely to become more prominent, complex, and deeply embedded in everyday life. 

Without strong policy frameworks in place, these systems may expand faster than governments and institutions can respond. 

While Moltbook may be perceived as “entertaining” on the surface, it reflects a broader transformation unfolding across the digital sector. 

Again, in this rapidly evolving landscape, the greatest challenge may not be artificial intelligence itself. 

It may be humanity’s ability to regulate it wisely.