> ## Content Index > Fetch the complete content index at: https://bytevyte.com/llms.txt > Use this file to discover other available public pages before exploring further. # AI Chip Controls and Safety Pledges Converge as Lawmakers Press Trump Before Xi Visit - URL: https://bytevyte.com/ai-chip-controls-and-safety-pledges-converge-as-lawmakers-press-trump-before-xi-visit/ - Published: 2026-09-18T04:15:16.000Z - Updated: 2026-09-18T04:15:16.000Z - Description: US lawmakers push Trump to merge AI chip controls with AI safety pledges before Xi Jinping's Washington visit, as Congress runs short on time to act. - Author: Bytevyte Editorial - Tags: ai-beats US lawmakers are pressing President Donald Trump to treat **AI chip controls** and the fate of detained Americans as one piece of leverage before Chinese President Xi Jinping arrives in Washington next month. Representative Ro Khanna has asked OpenAI and Anthropic to testify on frontier-model safety, and he says neither lab has committed despite repeated requests. He has also written to the Chinese developers Moonshot AI, Alibaba and DeepSeek, seeking binding commitments to a shared safety agreement with American AI laboratories. Those two tracks now run together. Export enforcement and frontier-AI rulemaking have become a single bargaining asset: Washington wants leverage over where advanced compute flows, and it wants both American and Chinese developers inside a framework it helps write. The letter to Chinese firms and the demand for testimony from US labs are two halves of one move. Bipartisan pressure has been building for weeks. Republican Congressman John Moolenaar, who chairs the House Select Committee on China, urged the administration earlier this month to ensure advanced logic chips do not reach Chinese firms considered untrusted. Senate Democratic leader Chuck Schumer wants a classified briefing for all senators on AI, plus guardrails that keep China away from US advanced chips. The detained-Americans file is older. Senior bipartisan lawmakers have warned about what they call hostage diplomacy and have urged Trump to raise the cases of political prisoners and unjustly detained Americans with Xi directly. Senator Bernie Sanders has separately pushed for AI safety talks with China. ## Why AI Chip Controls Are Now a Bargaining Chip Inside the administration, the answer is unresolved. White House AI adviser David Sacks persuaded Trump in May to scrap a planned executive order that would have put AI models through an extended government review, a reversal that caught senior officials off guard. The order's removal did not settle the question. It moved the fight to Congress, where no committee has produced a bill that can pass. The scrapped order matters beyond the May news cycle. A review regime would have created a standing process for government scrutiny of model releases, with timelines and disclosure requirements that labs could plan around. Its removal left voluntary practice and congressional pressure as the only constraints, and voluntary practice has no enforcement arm. Trump has stated flat opposition to new AI guardrails, arguing that regulation hands the race to Beijing. His position is that whoever wins AI wins everything, and that what he describes as a conspiracy against AI and data centers benefits only China. Congressional Republicans hold majorities in both chambers and have shown little appetite to break from him. House Republican leaders discussed AI regulation in a closed-door meeting on Monday and reached no decisions. Top White House officials are due to meet AI executives in the coming days. No plans have been announced to keep lawmakers in session to negotiate legislation, despite Democratic demands that they do so. ## Where the Four Camps Stand | Actor | Core ask | Instrument | | ------------------------ | ----------------------------------------------------------------------------------------------- | ----------------------------------- | | Rep. Ro Khanna | Testimony from OpenAI and Anthropic; binding safety pact with Moonshot AI, Alibaba and DeepSeek | Congressional letters and hearings | | Rep. John Moolenaar | Stop advanced logic chips reaching untrusted Chinese firms | House Select Committee on China | | Sen. Chuck Schumer | Classified all-senators briefing; guardrails; block China's chip access | Senate oversight | | David Sacks, White House | Scrap extended model review; retain chip leverage | Executive action and adviser access | | President Trump | No new AI guardrails | Veto and party discipline | Democratic House Minority Leader Hakeem Jeffries has argued for decisive action by lawmakers, and members of Congress spent this week meeting AI researchers and safety advocates. Any bill that reaches Trump's desk still needs his signature, and he has been explicit that he will not slow American AI progress. Schumer's demand for a classified, all-senators briefing is a procedural move with a substantive edge. It would put the administration's own assessment of Chinese model capability on the record for all 100 senators, several of whom will vote on export-control funding and on any eventual AI bill. Timing works against the regulators. The House is scheduled to adjourn within days until after November's midterm elections, which leaves a narrow window for AI legislation to move at all. ## The Trade-offs and the Clock Strict AI chip controls raise the cost of Chinese compute, but the enforcement record is thin. In July, lawmakers pressed the administration to close what they described as a loophole it had created in China tech restrictions, a gap that a key agency denied existed. Rules of that kind are hard to police and easy to route around through third countries, which leaves chipmakers and cloud providers planning against contradictory signals. Chips give Washington its sharpest instrument, and leverage of that kind decays. Every restriction creates an incentive to build substitute capacity, and the July dispute over the alleged loophole showed how quickly the line between enforcement and confusion blurs. A safety agreement would create a durable channel instead, which is why lawmakers keep tying the two issues together rather than trading one for the other. A binding safety agreement between American and Chinese labs would be the first of its kind, and no verification mechanism exists for it yet. Leahy has argued that superintelligence belongs on the agenda as a core national security threat, with mutual monitoring under a verifiable international regime so that neither side builds machines able to improve themselves without limit. What a bargain would require is easier to describe than to achieve. On the chip side, the ask is enforcement: keep advanced logic chips away from firms the US designates as untrusted, and close the routes that let them arrive anyway. On the safety side, the ask is a written commitment from Chinese developers to coordinate with American labs, which would be unprecedented in a sector where model weights and training methods stay closely held. The Chinese labs named in Khanna's letter have commercial reasons to refuse. DeepSeek, Alibaba and Moonshot AI compete directly with the American labs that would be their counterparties, and any monitoring regime would expose training decisions they treat as proprietary. Refusal costs them little in Washington, where the administration is already leaning on export controls rather than on agreements. The American labs face an inverted calculus. Testifying in public invites questions about safety claims that have not been independently verified, while staying away invites legislation drafted without industry input. Neither option is comfortable, which is why neither OpenAI nor Anthropic has committed. The White House session with AI executives is the likeliest venue for anything concrete before the summit. If the administration extracts a voluntary safety pledge from the labs, it can present that to Xi as evidence that American industry is already coordinating, without conceding a regulatory regime it opposes. Midterm politics shape the timeline more than the substance does. Members in both parties want a record on AI before November, while the administration prefers executive leverage, where it controls the pace, to legislation, where it does not. That mismatch explains why this month's loudest proposals arrived as letters and briefing requests rather than as bills with committee markups behind them. The detention request is not new. A similar bipartisan appeal reached Trump in May, ahead of an earlier round of talks, and the issue has returned to the agenda as the next meeting approaches. That repetition matters: lawmakers are using the summit calendar, not the legislative calendar, to force movement on consular cases. ## Why this matters For companies buying compute, the near-term variable is whether AI chip controls tighten or loosen in the next quarter, because export rules set the supply curve for training capacity and the price of inference. For frontier labs, the risk runs the other way: stay silent and inherit a standards regime drafted without them. Xi's visit forces chip enforcement and safety rules onto the same page, and Congress adjourns before it can answer either question on its own. ## Related Articles - [US and China Launch High-Level AI Regulation Dialogue to Set Safety Standards](https://bytevyte.com/us-and-china-launch-high-level-ai-regulation-dialogue-to-set-safety-standards/) - [China AI Export Controls Under Review as Beijing Tightens Tech Restrictions](https://bytevyte.com/china-ai-export-controls-under-review-as-beijing-tightens-tech-restrictions/) - [Trump Dismisses AI Safety Regulation as Hoax, Opts for CEO Talks Over Lawmaking \[Update\]](https://bytevyte.com/trump-dismisses-ai-safety-regulation-as-hoax-opts-for-ceo-talks-over-lawmaking-update/) ✔Human Verified --- *Researched and cross-referenced against primary sources by the Bytevyte editorial team. This article was generated with the assistance of artificial intelligence and reviewed by the Bytevyte editorial team.*