Source: U.S. House of Representatives (sito del deputato Gottheimer)

TEANECK, NJ, September 18, 2026: U.S. Congressman Josh Gottheimer announced a bipartisan AI safety agenda that includes two bills and an inquiry into the Administration’s AI cybersecurity plan. The American AI Security Act would establish a mandatory, pre-deployment national security review for the most powerful AI models, require developers to give the National Security Agency full access to evaluate risks such as cyberattacks or facilitation of chemical, biological, or radiological weapons, and operate on a 30-day review clock with a single 30-day extension. Companies could provide technical support to reviewers and appeal decisions through an expedited process modeled on courts used for national security cases.

The package also includes the China FIREWALL Act, which would ban Chinese-developed open-weight AI models from government-issued devices and prohibit federal agencies from procuring software that relies on those models, building on Gottheimer’s earlier No DeepSeek provision in the FY26 NDAA. Gottheimer said he is sending a letter to the President’s Cyber Director about the Administration’s AI cybersecurity clearinghouse. The release cites recent incidents involving Hugging Face and disclosures about autonomous AI agents as motivating concerns, and the announcement was made at Fairleigh Dickinson University during the launch of its Health Humanics Innovation Hub; attendees included Senator Gordon Johnson, Teaneck Mayor Mark Schwartz, FDU officials, and students.