The administration’s emphasis on winning an "AI race" has coincided with a series of measures that limit Chinese access to advanced chips and software. Export controls have been tightened, and Congress is considering legislation that would further restrict hardware shipments. In the past fortnight, the Justice Department charged six leading Chinese AI firms with alleged theft from U.S. frontier laboratories, a claim Beijing has vigorously denied.
Several U.S.,China specialists told Ars Technica that the heightened rhetoric could prompt Beijing to walk away from dialogue, yet public Chinese reactions have been muted. Even a post on the former president’s Truth Social platform, in which he declared that whoever wins AI will win overall, did not deter Chinese officials from attending scheduled meetings, according to the experts.
A proposed bilateral alert system aims to transmit AI safety warnings, but critics note the plan lacks technical expertise on either side. Sam Bresnick, a research fellow at Georgetown’s Center for Security and Emerging Technology, told Ars that the existing channel is preferable to creating a new one, but he expressed limited confidence that it will substantially improve risk management.
Su Lian Jye, chief analyst at Omdia, told the South China Morning Post that the alerts could increase transparency but would not erase underlying mistrust. He argued that the mechanism would only become effective if both nations agreed on risk classifications, trigger thresholds, response procedures, and periodic reviews, emphasizing that without such details the dialogue would remain superficial.
The issue is expected to surface during the upcoming meeting between President Trump and President Xi, where AI is slated as a top agenda item. On CBS’s Face the Nation, Representative Ro Khanna, the ranking Democrat on the House Select Committee on Strategic Competition, urged that talks go beyond a simple line between officials lacking technical backgrounds, calling for deeper cooperation on safety protocols.
Alvin Wang Graylin, a global technology strategist at Stanford Digital Economy Lab, told Ars that the alert proposal represents a surprisingly positive development from recent negotiations. He highlighted that China routinely audits AI models and could share insights from hundreds of reviews, while the United States is only beginning to conduct systematic audits, suggesting a mutually beneficial exchange of risk-mitigation practices.