Two US lawmakers have introduced legislation calling for binding and verifiable artificial intelligence (AI) safety safeguards with China, including a direct government crisis hotline between Washington and Beijing to manage serious AI incidents. The proposal, tabled on 23 September 2026, represents one of the most structured legislative attempts yet to frame US-China AI risk coordination as a national security imperative.

What the Legislation Proposes

Representatives Sam Liccardo and Kevin Kiley, both of California, said the bill would promote coordination between the two countries on risks from advanced, or frontier, AI systems. The framework would remove existing legal barriers that currently prevent American AI safety experts from engaging in narrowly defined technical discussions with their Chinese counterparts.

It would direct the State Department and other parts of the administration to negotiate safeguards covering model testing, independent audits, transparency, incident reporting, and the verification of compliance. Monthly updates to Congress from the State and Commerce departments would also be required under the proposed framework.

The Case for Engagement, Not Trust

Representative Liccardo was direct about the rationale. 'A lot of technical work must begin now, before political leadership can ever fruitfully meet at the negotiating table,' he said. He drew a pointed historical comparison: 'We don’t talk to the Chinese because we trust them or agree with them; we shouldn’t and don’t. We talk to the Chinese because — as with the USSR during the Cold War — all of our lives depend on it.'

Liccardo also pushed back on the prevailing AI industry ethos of speed above all else. 'The AI industry has focused on running faster, but no human wins a race to superintelligence that exceeds human control,' he said, adding that establishing 'red lines now' could 'avoid the need for red phones later.'

Key Safeguard Areas and IP Protections

The framework would focus negotiations on risks involving the loss of human control over AI systems, failures in AI alignment, AI-enabled chemical, biological, radiological and nuclear (CBRN) threats, and major cybersecurity dangers. Routine technical communication between specialists from both nations would be mandated to prevent dangerous AI incidents from escalating into broader geopolitical crises.

The bill would also instruct the Commerce Department to pursue protections against compromise of American intellectual property through model distillation — a process that generally involves transferring capabilities from a larger AI model to a smaller system. US policymakers have raised concerns that the process could be exploited to reproduce advanced American AI technology. Participants in any joint forum would be explicitly prohibited from sharing proprietary intellectual property, distillation-enabling information, or sensitive national security data.

Funding and Institutional Framework

The legislation would authorise $100 million for the Center for AI Standards and Innovation, which would bring together experts from industry, universities, and civil society to develop common testing, evaluation, and verification standards. Both Liccardo and Kiley framed the bill as a pre-emptive architecture — building the technical and institutional scaffolding before diplomatic negotiations can begin in earnest.

With frontier AI development accelerating on both sides of the Pacific, the bill's proponents argue that structured engagement — however uncomfortable — is less risky than no engagement at all. Whether Congress advances the proposal will be a key indicator of how seriously Washington is treating AI governance as a strategic priority.