Simulating the simulation

How AI is a reflection of our own social behavior

SocialAI.co is a subtle yet striking commentary on the Zeitgeist. At first glance, it might seem indistinguishable from any other social network—and that’s the point. But here’s the twist: all your followers are AI chatbots. Every post you make is seen and responded to by these AI personas, each designed to engage, challenge, or even entertain you.

It’s a platform that mirrors human interaction without the unpredictability of real humans—or rather, with a different kind of unpredictability. Like Twitter or Threads, you never quite know when humor will fall flat or when meaning will be lost. You only know that it will.

This brought me back to a scenario generator I built with Johannes Kleske and Fabian Mürmann using the GPT-3 API (2021!). You’d enter a keyword and receive three randomized future scenarios. Most were bizarre or irrelevant, but occasionally you’d stumble upon something profound—fuel for new ideas. Even then, it was clear how this new version of creativity could be useful, wasteful, fascinating, and scary, all at once.

SocialAI echoes this same dynamic. It exposes how transactional social media has become—sometimes uncomfortable, yet full of surprises. I can imagine social media teams using it to simulate responses and predict the outcomes of certain messaging. While it’s far from perfect, some simulation is better than none.

For everyday users, the possibilities are just as interesting. Could this be a new way to create memes, where the AI generates responses, or could people test the potential virality of their ideas? The novelty lies in how people will adapt and push the boundaries of this AI-only social world.