EU AI rules highlight gap in US oversight

The EU Artificial Intelligence Act, in force since August 2024, is rolling out phased rules intended to make AI systems more transparent, accountable and easier to challenge. A centralized database planned for December 2027 will require certain high-risk systems to be registered with details on providers, intended uses and deployment, and consumers will be able to file complaints if required systems are missing.

Transparency obligations that began on August 2nd 2026 require companies to disclose certain AI interactions, including customer support bots and other interactive systems, rather than presenting them as human. The rules also cover deepfakes and AI-generated images, video and audio, while providers of generative AI systems must support detectability and provide more documentation on models, copyright policies and training content.

The law bans some AI uses involving manipulation, exploitation of vulnerable groups and unfair scoring. High-risk hiring tools are subject to oversight, registration and eventual explanation rights for affected people. Employers also must ensure workers using AI understand the systems, their limits and risks.

The US approach remains mostly state-based. Colorado attempted a broader model in 2024, but enforcement was paused in April of 2026 and later rewritten. Illinois, California and NYC’s Local Law 144 address employment disclosure, bias testing and hiring audits within their borders, leaving no comparable federal framework.