Everyone is trying to get in on the game, and the result is an unhelpful patchwork of laws, like what we see for data breach notices.

Through its National Policy Framework for Artificial Intelligence released earlier this year, the White House is telling states and asking Congress to tell states to stop regulating AI. Despite this effort, states are powering forward.

Bret S. Cohen, a data privacy partner at Hogan Lovells Cadwalader, told Law 360 that “[e]veryone’s trying to understand how AI can be used and what the benefits are, so what we’re seeing is more of this harms-based, Whac-a-Mole-style legislation, which is likely to continue as more specific harms coming out of the development or use of AI are identified and targeted by policymakers.”

Everyone is trying to get in on the game, and the result is an unhelpful patchwork of laws, like what we see for data breach notices.

 For example, in May, Connecticut passed a landmark artificial intelligence law. This sweeping legislation regulates automated employment decisions, AI companion safety protocols, generative content provenance, and youth online protections. In Illinois, a groundbreaking law established a framework for AI safety, transparency, and accountability for AI models. Under this law, the Artificial Intelligence Safety Measures Act (AISMA), Illinois is the third state to require frontier model developer transparency and the first state to require third-party audits of artificial intelligence models.

 Receiving lots of media attention is Colorado. The state replaced an earlier and substantial regime to tackle potential bias when AI systems are used in making employment and other consequential decisions with a narrower set of restrictions imposing notification requirements on companies that use AI in making “consequential decisions” impacting workers or consumers.

 Congress has quite a few bills before it dealing with AI governance and the FTC is seeking public comment on a proposed policy statement addressing concerns that AI companies may be manipulating the behavior of their AI systems contrary to reasonable consumer expectations for objectivity and accuracy.