Brussels has become the geographic centre of a widening argument over how AI should be governed. As of August, key provisions of the EU's AI Act are now legally enforceable, requiring companies deploying high-risk AI systems in the bloc to meet transparency, risk-assessment and human-oversight obligations or face fines. Washington, meanwhile, has been pushing the opposite direction, favouring lighter-touch rules that officials argue keep American AI companies competitive against faster-moving rivals.
The divide is being described in policy circles as a potential unwinding of the so-called "Brussels effect," the pattern by which EU regulation has historically set a de facto global standard that companies adopt everywhere rather than build separate compliance regimes per market. With the US actively courting AI firms with a looser rulebook, some companies now face a genuine choice between two regulatory philosophies rather than a single global baseline set by European law.
European officials in Brussels have shown little sign of softening the bloc's approach in response. The Commission's position, reiterated as the AI Act's obligations took effect, is that binding rules around transparency and human oversight are what will make AI systems trustworthy enough for wide adoption in Europe, even if that means some AI companies choose to launch new products in the US first and delay a European rollout to manage compliance costs.
