Rep. Josh Gottheimer (D-N.J.), one of three co-chairs of the House Democratic Commission on Artificial Intelligence, said Friday he intends to introduce two bipartisan bills that would address surging AI safety concerns.
The New Jersey Democrat — who along with his fellow co-chairs was tapped by House Minority Leader Hakeem Jeffries to chart his party’s path on AI — is likely to play a key role in setting the congressional agenda for the technology should Democrats retake control of the House in November’s midterm elections.
At an event held at Fairleigh Dickinson University, Gottheimer announced the American AI Security Act, legislation with Rep. Mike Lawler (R-N.Y.), that would create a mandatory government review process for advanced AI models ahead of their release. The reviews would be led by the National Security Agency and would last for 30 days, with the possibility for a single 30-day extension.
“Right now, there are no mandatory requirements for pre-deployment testing of frontier AI models,” Gottheimer said. “That doesn’t work in banking, it doesn’t work in food safety, it doesn’t work in medicine and it sure as hell shouldn’t work with the technology that experts say could help build a bioweapon or take down our power grid.”
It’s an echo of Gottheimer’s statement earlier this week, where he poured cold water on OpenAI’s endorsement of a provision in the bipartisan FRONTIER Act that would require third-party safety assessors to be embedded within leading AI firms. He suggested the companies should be forced to go further and submit their models directly to the government for review.
Gottheimer’s plan broadly follows the Trump administration’s existing vetting regime for advanced AI models, which came together in chaotic fashion this summer in response to cybersecurity risks raised by new models from OpenAI and Anthropic. The White House has kept specifics of that process under wraps.
The New Jersey Democrat also pledged to soon introduce the China FIREWALL Act, legislation with Rep. Nick LaLota (R-N.Y.) that would ban Chinese-developed AI models with open weights from government-issued devices. It would also block federal agencies from procuring software that is reliant on those models.
Read the full article here