OpenAI executive urges caution in China-US AI talks, warns of distillation threats
A top OpenAI executive said on Monday that Washington needs to be “clear-eyed” and “aware” about sharing information with Beijing during coming talks on artificial intelligence (AI), while also warning about the threat posed by alleged Chinese distillation attacks. “Be very clear-eyed. Be very aware of what information you are able to expect in return for what you are sharing, and start to build…
OpenAI's deputy lead for national security, Caroline Zier, warned on Monday that Washington needs to be cautious when sharing information with China during upcoming talks on artificial intelligence (AI). Zier, who previously served as a special assistant to the US secretary of defense, emphasized the need to be "very clear-eyed" about the information exchange and "build a foundation" based on mutual understanding of risks.
She expressed skepticism about the alignment of US and Chinese perspectives on AI safety and loss of control during the talks. Zier also addressed the threat of unauthorized distillation attacks, which involve malicious actors using a "teacher" model's outputs to train a "student" model without authorization. These attacks, often originating from Chinese labs, aim to bypass safety guardrails and proprietary intellectual property.
US Treasury Secretary Scott Bessent and Chinese Vice-Premier He Lifeng are set to meet before the upcoming summit between US President Donald Trump and Chinese leader Xi Jinping, where a potential AI cooperation framework is expected to be discussed. Despite the potential benefits of AI collaboration, both experts caution against becoming overly dependent on Chinese models, citing the potential risks of restricted access and increased costs.
Written by urgent.news from Reuters Business via SCMP's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
