The US will put AI-directed cyberattacks on the table when Trump meets Xi this month.
Nikkei Asia reported Sunday that the United States plans to raise AI safety, specifically AI-directed cyberattacks, when Xi Jinping visits the White House on September 24. The report cites people familiar with the matter. It comes after two months of stories about rogue AI agents being used for hacking, stories that started as isolated incidents and have apparently moved fast enough to land on a head-of-state agenda. Nikkei's report did not detail what the US will specifically propose or how Beijing is expected to respond.
That escalation path is the real story here. Cybersecurity has been a recurring irritant in US-China relations for over a decade, but autonomous AI-driven attacks are new enough that there is no existing diplomatic playbook for them. Putting the topic in front of Trump and Xi directly, rather than leaving it to working-level technical talks, suggests Washington sees this as urgent rather than routine.
Past US-China cyber dialogues have mostly produced statements of concern rather than enforceable limits, so the more interesting question is whether this one ends differently.