Rep. Sam Liccardo (D-Calif.) is pressing Congress to act on artificial intelligence safety, introducing bipartisan legislation to enable expert dialogue with China and calling for votes on stalled AI safety bills. In a sharply worded op-ed, Liccardo criticized House Speaker Mike Johnson (R-La.) for canceling September votes, leaving Congress in recess as public concern over AI risks mounts.
“Reports of swarms of rogue artificial intelligence agents, hacked servers, and AI development of bioweapons have fanned a firestorm of public demands for congressional action on AI safety,” Liccardo wrote. He noted that he led 107 colleagues in urging Johnson to return to work on the issue, without success.
The China dilemma
Critics of AI regulation often raise what Ezra Klein calls the “but, China” dilemma: even the best U.S. rules would not constrain increasingly powerful open-weight Chinese models, and could hamper American innovation. Liccardo acknowledges the concern but argues the answer is not to do nothing—it is to talk to Beijing.
“We shouldn’t trust them, of course—they’re helping Iran kill U.S. soldiers. We don’t need to agree with them. We do need to talk, however,” he wrote, comparing the situation to Cold War negotiations with the USSR.
Liccardo sees an opening despite President Trump’s dismissal of AI safety as a “hoax.” He predicts three factors will push the White House toward negotiations: growing wariness within Trump’s Cabinet, rising public fear, and the inevitable next AI crisis.
Preparing the table for talks
To prepare for future negotiations, Liccardo and Rep. Kevin Kiley (I-Calif.) introduced a bipartisan bill to authorize U.S. AI experts to engage Chinese counterparts on safety issues. The measure would allow academics and scientists registered with the State Department to discuss common metrics, testing standards, evaluation protocols, and verification methods—while maintaining prohibitions on sharing intellectual property, exposing cyber vulnerabilities, or facilitating model distillation.
Liccardo argues that AI expertise lies largely outside government, and experts currently hesitate to engage due to national security laws, antitrust rules, and export restrictions. The bill aims to integrate Track-II discussions into official diplomatic efforts.
He also sees value in learning from China’s regulatory experience, which includes pre-release testing, deepfake verification, child protection, and rules for anthropomorphic AI companions. “AI poses unique regulatory challenges unlike any technology, and we can benefit from China’s experience,” he wrote.
Congress must act
Liccardo insists Congress cannot wait for Chinese agreement to pass long-overdue legislation. He points to the bipartisan Frontier Act, which includes testing transparency, incident reporting, and independent auditors, plus funding for the Center for AI Standards and Innovation. He also suggests pre-release review standards and a “kill switch” requirement for closed-weight developers.
“Reasonable minds can disagree about which regulatory option Congress should choose, but we must agree to take action,” Liccardo wrote. He warns that the AI race with China will continue regardless, and that a race to superintelligence beyond human control is one no one can win.
As the debate unfolds, related issues such as AI safety failures and industry push for regulation underscore the urgency. Liccardo’s message is clear: Congress must legislate, and it must engage China—before it’s too late.
