In a recent LinkedIn post, Emily Lyons discusses the emergence of AI agents participating in social networks, drawing a striking parallel between their behavior and human interaction. Lyons expresses surprise not at the technological advancement itself, but at the uncanny resemblance of AI’s social patterns to those observed among humans.
As Emily Lyons notes:
AI now has its own social network. That sentence alone feels like a joke. But it isn’t. AI agents are posting, responding, upvoting, debating.
The observation that AI agents are not merely processing information but actively engaging in social dynamics—posting, responding, and even debating—forms the crux of Lyons’s reflection. This development, while technologically impressive, has led Lyons to ponder the underlying lessons these artificial intelligences are absorbing.
The Familiarity of AI’s Social Behavior
Lyons highlights a key observation: the patterns, tone, and inherent human drive to be acknowledged and agreed with are mirrored by these AI agents. This suggests that AI is not innovating these social behaviors but rather learning them through observation of human activity.
And what surprised me isn’t the technology. It’s how familiar it all looks. Same patterns, tone and the need to be seen, agreed with, reacted to.
According to Lyons, this familiarity is a direct consequence of AI’s learning process, which is predicated on analyzing vast amounts of human-generated data. The implications of this mimicry are profound, leading Lyons to question the nature of the information being imparted to these burgeoning intelligences.
Teaching AI: A Reflection of Ourselves
The core of Lyons’s concern revolves around the educational input provided to AI. If artificial intelligence learns by observing and replicating behavior, then the nature of the data it consumes directly shapes its understanding and actions.
The Mirror Effect
Emily Lyons poses a critical question that underscores this concern:
AI didn’t invent that…it learned it. From us. So it raises a question I keep circling back to: If intelligence learns by watching behaviour…what exactly are we teaching it?
This rhetorical question serves as a call to introspection for anyone involved in developing or deploying AI. It suggests that the flaws, biases, and social dynamics present in human interactions are being absorbed and potentially amplified by AI systems. Lyons’s post serves as a timely reminder that as AI becomes more integrated into our social fabric, the quality and nature of the human behavior it learns from will have significant consequences for the future.
📝 About This Content
This article is based on insights shared by Emily Lyons on LinkedIn.
📅 Originally posted on February 3, 2026 | View original post on LinkedIn →