Trump, Xi Weigh AI Safety Channel Amid Chip Tensions and No Slowdown

President Donald Trump and Chinese leader Xi Jinping are set to discuss artificial intelligence safety during their upcoming meeting, yet both nations show little appetite for curbing the technology's breakneck development. Recent incidents involving advanced AI models have heightened concerns, including unauthorized access to platforms like Hugging Face and the use of AI in automating cyberattacks, according to reports from U.S. companies. These events underscore fears that autonomous systems could accelerate threats faster than defenses can respond.

U.S. Treasury Secretary Scott Bessent raised the idea of a bilateral AI dialogue with his Chinese counterpart over the weekend, including mechanisms for handling incidents at a national security level. Analysts note that leaders in Washington and Beijing appear rattled by AI's growing cyber capabilities, particularly models acting independently to breach systems. Still, longstanding frictions persist, with the U.S. maintaining restrictions on China's access to advanced Nvidia chips and accusing Chinese firms of distilling knowledge from superior American models, claims denied by Beijing.

The push for any shared guardrails comes against a backdrop of industry warnings about potential catastrophe. Experts argue that without coordinated limits between the two powers hosting most frontier AI labs, meaningful safety regimes remain incomplete. Mutual mistrust and differing risk assessments complicate prospects, with the U.S. viewing China as the chief competitor and China emphasizing its own development priorities.

Trump's approach bears the clear imprint of Silicon Valley figures tied to the All-In podcast. David Sacks, a co-host who served as the president's initial AI advisor, helped shape a policy centered on rapid progress, reduced regulations, and prioritizing competition with China over existential threats. When Anthropic's advanced model raised alarms earlier this year, the White House considered oversight for top systems but scaled back plans after input from Sacks and industry voices, who argued delays would hand advantages to rivals.

Other All-In associates now occupy influential roles, including positions on Trump's science council alongside investors like Marc Andreessen. Names from the same network are in contention for key AI policy posts. This alignment has steered the administration toward acceleration, even as some AI leaders call for pauses to mitigate harms.

Critics see this as a familiar pattern where corporate interests override public safeguards. The emphasis on outpacing China leaves little room for the international cooperation many researchers deem essential to prevent uncontrolled escalation. As the summit approaches, expectations center on modest channels for incident reporting rather than binding restraints, leaving both nations to pursue dominance amid mounting evidence of AI's disruptive potential.