OpenAI national security deputy Caroline Zier backed US-China dialogue on artificial intelligence, but argued that engagement should be measured: Washington should consider what information it shares, what it can reasonably expect in return, and whether the two countries understand AI risks in the same way. She also raised concerns about alleged Chinese AI-model distillation.
10
12
Dialogue, with realistic expectations
Speaking at a Hudson Institute discussion on November 2, 2026, Zier said information-sharing should be approached “clear-eyed” and with an understanding of what the United States might receive in return. She called for building a foundation for engagement, rather than treating talks as a reason to share without limits.
1
10
Her position was not a rejection of bilateral dialogue. The reporting describes her as supporting engagement with Beijing while emphasizing that expectations about reciprocity should remain realistic.
12
Don’t assume a shared view of AI risks
Zier also questioned whether the United States and China mean the same thing when they discuss AI risks. That caution matters to the terms of any dialogue: conversations about safety or risk do not, by themselves, establish a common understanding.
12
Distillation remains a security concern
Alongside her support for talks, Zier warned about alleged Chinese distillation attacks. The reports frame these as a threat to US AI technology; the allegation should not be treated as independently established fact based on the available reporting.
10
The overall message was to pursue engagement without confusing it with trust: build a basis for dialogue, be deliberate about disclosures, and keep technology-security concerns in view. The available sources report her remarks but do not provide a full transcript for checking every detail of the discussion.
6
10