AI regulation: the US fragments while the EU remains (for now) a single market

In early August 2026, Illinois enacted the first US state law requiring mandatory audits of so-called "frontier" AI systems, with civil penalties of up to $3M per violation. Colorado, for its part, strengthened its chatbot-specific obligations. At the federal level, the GAAIA (Government Advancing Artificial Intelligence Act) preemption bill, meant to harmonize these disparate state rules, remains stalled in Congress, according to law firm Mintz's "AI: The Washington Report" (August 2026 edition).

For an industrial or luxury company deploying AI systems across multiple US states, this gridlock translates into a stack of local rules to satisfy simultaneously — mandatory audits in Illinois, conversational transparency obligations in Colorado, and a patchwork set to grow as more states legislate in turn.

The comparison with the European Union is instructive. The AI Act, despite its own repeated schedule slippages, remains a single text applicable across all 27 member states, with fines capped at €35M or 7% of global turnover for prohibited practices — roughly ten times Illinois's cap in absolute terms, but within a unified framework rather than a fragmented one.

The takeaway for aerospace, energy, or luxury groups operating in the US isn't to wait for a hypothetical federal preemption law: state-by-state obligation mapping is becoming an AI governance prerequisite, regardless of how the European AI Act's timeline plays out.