In a move to preempt potential dangers associated with the rapid advancement of artificial intelligence (AI), Representative Josh Gottheimer (D-N.J.) announced his plan to introduce two significant bipartisan bills aimed at tightening AI safety protocols. As co-chair of the House Democratic Commission on Artificial Intelligence, Gottheimer’s initiative underscores a proactive approach to navigating the complexities of AI development and deployment.

During a recent gathering at Fairleigh Dickinson University, Gottheimer, alongside Rep. Mike Lawler (R-N.Y.), unveiled the American AI Security Act. This proposed legislation mandates a comprehensive government-led review process for new AI models before they hit the market. The National Security Agency would spearhead these evaluations, which could extend up to 60 days. Highlighting the absence of stringent pre-deployment testing in the AI sector, Gottheimer drew parallels to established safety standards in banking, medicine, and food safety, stressing the importance of rigorous vetting to prevent misuse of AI technologies in scenarios as severe as bioweapon production or power grid sabotage.

This legislative effort mirrors sentiments expressed earlier in the week when Gottheimer critiqued OpenAI’s support for the FRONTIER Act's proposal for third-party safety assessments within AI firms. He advocated for a more stringent review process, suggesting direct governmental oversight rather than relying solely on internal corporate evaluations.

The Trump administration's recent approach to AI model vetting, albeit formulated under chaotic circumstances, also influenced Gottheimer’s proposed framework. This existing policy, which largely remains shrouded in secrecy, was a response to heightened cybersecurity threats identified in new AI developments by companies like OpenAI and Anthropic.

Additionally, Gottheimer plans to introduce the China FIREWALL Act with Rep. Nick LaLota (R-N.Y.). This bill aims to safeguard national security by prohibiting Chinese-developed AI models with open weights on government devices and blocking federal procurements dependent on such technology.

These legislative endeavors by Gottheimer not only highlight the bipartisan recognition of the risks associated with AI but also signal a significant shift towards establishing a controlled and secure framework for AI development in the United States. As AI continues to integrate into various sectors of society, these proposed laws could serve as critical benchmarks for future AI governance, ensuring both innovation and safety are prioritized in the digital age.