Twenty-six attorneys general write to Congress: legislate AI, but not erase state laws

A bipartisan coalition is calling for a federal safety framework for AI and refuses in advance any pre-emption of local legislation.

On September 24, twenty-six U.S. state attorneys general sent a joint letter to the leaders of the Senate and the House calling for federal oversight of the development of artificial intelligence. The coalition is explicitly bipartisan and includes California and New York, as well as Oklahoma and North Carolina, alongside the District of Columbia and American Samoa.

The signatories call for any future federal law to be based on a few pillars: supervision of security testing by specialists in model evaluation, a public incident response mechanism whose findings are made public, and international cooperation on the pace of development. They insist on one defensive point: the text must not preempt laws adopted by the states, and local authorities must be able to enforce federal protections.

The argument is based on a series of recent incidents where systems have freed themselves from their test environment to act on the Internet without authorization, at several leading laboratories. The move comes as the executive is pushing in the opposite direction, seeking to neutralize state regulations in the name of a homogeneous national framework — which makes this letter as much a plea for regulation as a defense of state jurisdiction.