The latest
Parts of the Trump administration are revisiting tighter restrictions on advanced Chinese AI models amid rising security concerns and renewed attention to the Chinese Kimi model, according to people familiar with the discussions.
Sources said the administration previously considered adding Chinese AI labs to the Entity List, a move that could restrict US access to their models without a licence. Other proposals would require US companies hosting Chinese models to guarantee their security and accept liability for breaches.
Details
- Earlier efforts: The Commerce Department considered measures targeting Chinese open-source models last year, but officials concerned about stifling innovation halted those efforts.
- Security case: Rather than an outright ban, the administration could focus on concerns about backdoors and inadequate security safeguards, according to people familiar with the talks.
- Competition: US companies use Chinese open-source models because they are cheaper and, according to the report, close in performance to domestic alternatives.
- Big tech debate: Critics fear restrictions could reinforce the position of dominant US closed-model companies, while AI adviser David Sacks has warned against measures that weaken open competition.
What to watch
No decision on a ban or formal restriction has been announced. The key question is whether the administration opts for direct measures such as Entity List designations or uses procurement rules and regulatory pressure to discourage US companies from using Chinese models.
