US and China discuss AI risks but neither wants to slow development

In short: The US and China are talking about the risks of advanced AI, but both are still pushing to move faster than the other.

What's going on

Recent reporting suggests US and Chinese leaders are expected to discuss safety concerns around artificial intelligence, including cyberattacks, scams, and the use of AI in weapons. There is no publicly confirmed 2026 summit focused only on AI risk, but the topic is showing up in side conversations at big international meetings.

At the same time, both countries are acting as if they are in a race. The US has tightened restrictions on selling some top AI chips to China. These chips are like the engines that power the most capable AI systems. US officials say limits are tied to national security worries, such as AI helping with cyberattacks or dangerous biology.

China is also pouring money into AI and has said it will manage risks mostly through its own rules, not by slowing down overall progress. Reports describe Chinese plans for safety checks on very large AI models, while continuing to expand AI use in both civilian and military areas.

Experts have warned for years that AI is moving faster than safety work. Some researchers have called on governments to spend much more on testing and safeguards, similar to how you would require stronger brakes as cars get faster.

What to watch

Watch for any joint statement that sets shared “guardrails” (basic rules and limits) on issues like cyberattacks, deepfakes (fake videos or audio), and autonomous weapons. Also watch whether either country backs up safety talk with budget and enforcement, not just meetings and announcements.

Source: NYTimes