AI and Local Deliberation for Better Rules

Laws were originally created to mediate conflicts among members of society. Yet most of the laws we live under today are not the product of deep deliberation. They are, more often than not, rushed compromises shaped by tight deadlines and political constraints. As legislation driven by schedules rather than careful reasoning continues, institutions drift further from the realities they are meant to govern.

Why does this pattern persist? Many attribute it to the limited capacity of legislators or short-term political interests. However, there is a more fundamental issue. Modern societies consist of millions of people holding vastly different values, lifestyles, and priorities. When a single, uniform set of rules is applied simultaneously across an entire nation, friction is inevitable somewhere in the system. What feels like a minor inconvenience to some can amount to serious harm for others. Moreover, it is nearly impossible to foresee, at the moment of legislation, what long-term consequences a seemingly minor provision might produce decades later.

This is where expectations around AI arise. But those expectations need to be recalibrated from the beginning. AI is not a substitute that can design perfect laws in our place. What it can do effectively is simulate outcomes: showing where a proposed policy is likely to create conflict, which groups—especially minorities—stand to be adversely affected, and what probable long-term results might follow. In short, AI is not a tool for replacing human judgment, but for providing better information on which to base that judgment. The final decision must remain with people.

If AI is to be genuinely useful, where should it be applied? Running simulations at the national level alone will not resolve the structural problems of centralized lawmaking. A more promising and realistic path lies in regional experimentation.

Rather than having the central government impose identical rules on every region, each area should be allowed to design and test policies suited to its own circumstances. Successful experiments can be adopted elsewhere, while those that produce negative consequences can be revised or discarded quickly. When multiple regions pursue different approaches simultaneously, the risk that the entire society suffers severe damage from a single flawed decision is significantly reduced. This model challenges the assumption that the center must design everything from above.

For such experimentation to generate real progress rather than mere fragmentation, a crucial mechanism must be in place: the movement of people. Over time, it becomes clearer which systems work better and which do not. Regions that cling to irrational or oppressive rules will likely experience population outflow, while those that develop practical and humane institutions will tend to attract people. This movement itself serves as a powerful feedback loop. Declining regions receive a clear signal that their rules require serious reconsideration.

However, this freedom to experiment must have a firm, non-negotiable boundary. Majority opinion must never be allowed to suppress minorities or excessively restrict fundamental individual rights. Bodily autonomy, freedom of expression, and protection from discrimination must remain protected as minimum constitutional guarantees. The state should maintain this essential framework while giving regions substantial room to experiment in other areas.

The most critical condition for this system to function properly is freedom of movement. When the norms or institutions of a particular region do not suit someone, they must have the real ability—both physically and institutionally—to move to another region or even form a new community. Without this guarantee, regional autonomy risks becoming nothing more than smaller-scale coercion. Only when people can genuinely choose their community does experimentation carry meaningful weight.

The centralized approach of trying to redesign every aspect of life from the top has already revealed its limits. AI can help mitigate those limits, but it cannot overcome them on its own. AI's role is to make regional experimentation safer and more informed. In the end, the question of what kind of society we want to build remains one that humans must answer for ourselves.