AI Can Simulate Society. But Can It Help to Transform It?
- Derya Yüksek
- 3 days ago
- 3 min read
AI-powered simulation is becoming increasingly capable of modeling collective dynamics, capturing how societies behave — but what would it mean to use it to think through transformation?
With agent-based models — and now large language model (LLM)-driven systems — we can simulate how opinions spread, how polarization emerges, and how collective dynamics evolve. Recent work on generative agents shows how AI-driven entities can mimic believable social behaviour, while newer research points to the growing role of LLMs in social simulation.
We can model complexity better than ever. But something is still missing.
Conflict Is Not Just Behaviour
When we talk about conflict, we often focus on what people do — who says what, who reacts how, what escalates, what de-escalates.
But conflict is also about how situations are understood and felt. It is shaped by:
how issues are framed
how emotions circulate
how groups relate to each other
and what people see as possible or legitimate
These dimensions — meaning, affect, and relationality — have been studied for decades. But they rarely sit at the center of how we model collective dynamics.
These dimensions are not separate. They constantly shape each other.
A narrative can intensify fear. Fear can harden boundaries. Boundaries can limit what kinds of responses feel acceptable.
Understanding this interplay is key — not just for analysis, but for thinking about transformation.
The Limits of Modeling
Today’s simulations are powerful. Still, most of them are still built to answer one core question:
What will happen?
They help us observe patterns. They help us understand behaviour.
But they are less equipped to engage with a different kind of question:
How might things be approached differently?
From Modeling to Exploration
So what if simulation wasn’t only about prediction? What if it could also become a space for exploration? Instead of asking “what will happen?”, we might ask:
What becomes possible if things are approached differently?
This is not about predicting outcomes.
It’s about working with existing research, practices, and lived experiences to explore how different ways of engaging with conflict might interact with real-world dynamics.
Sometimes that means changing how an issue is framed. Sometimes it means creating new spaces for dialogue. Sometimes it means shifting relationships — or how they are perceived.
No single approach “solves” conflict.
But different approaches can reshape how situations unfold.
The Missing Piece: Reflexivity
There’s another layer that is often overlooked.
Most simulation systems place the user outside the system — as an observer.But in reality, we are never outside.
The way we interpret a situation shapes what we see. What we see shapes what we think is possible.
Bringing reflexivity into simulation means acknowledging this.
It means not only exploring dynamics, but also questioning the perspective from which we engage with them.
Why This Matters
We are living through deeply polarized and complex forms of conflict.
Understanding these dynamics is important. But understanding alone is not enough.
We also need ways to engage with them — carefully, thoughtfully, and across different layers.
This is not just a technical challenge.It requires collaboration between fields:
AI and computational modeling
communication and media studies
conflict transformation
and grounded, practice-based knowledge.
A Different Role for Simulation
One emerging direction is to rethink simulation as something more than a modeling tool.
Not as a system that tells us what will happen, but as a space that helps us explore how different approaches might interact with complex realities.
This kind of work is still developing.
And it reflects a broader shift: from prediction to exploration.
Not Predicting the Future — Expanding It
Simulation has already changed how we understand collective behaviour.
Its next step may be quieter, but more demanding: to help us think differently about what might become possible.
Comments