Trend Insight 18: jurisdictions diverge on what AI regulation is even for

Summary

This trend describes governments splitting not just on how strictly to regulate AI but on what regulation should achieve. Some jurisdictions write rules that define forbidden uses and assign liability for harm, as the European Union does by holding developers and deployers responsible. Others, such as California, focus on transparency, requiring disclosure of training data and watermarking of generated content. At the same time, lawsuits over copyright and consumer protection are testing whether existing legal categories fit AI at all, including whether an AI offering is a product or a service. Preventing misuse is also technically hard, as users keep finding ways around new guardrails. The result may be a fragmented global landscape in which companies release region-specific versions or leave the strictest markets.