The U.S. says China's AI progress is down to 'distillation.' But is it that clear cut?

The US administration and some AI developers, including Anthropic, have accused Chinese AI companies of using "distillation" to advance their AI capabilities. However, some voices are questioning the impact of this alleged practice. Distillation involves training smaller language models using prompts and responses from more advanced models, allowing them to emulate the capabilities of the more intelligent model.

According to Tom's Hardware, the US government and American AI developers are concerned that distillation attacks may be helping China and Russia develop AI models with similar capabilities at a lower cost and with fewer compute requirements. China has rejected these claims, but has pledged to take countermeasures if the US uses these allegations to restrict Chinese AI development.

Efforts to combat distillation attacks have been ongoing, with major Western AI labs pledging to work together to detect and prevent them. However, foreign actors have also been purchasing logs of third-party conversations made using legitimate accounts, making it challenging to halt the practice entirely.

Brief written by urgent.news from CNBC World, Tom's Hardware — 2 reports on this story. Machine-written — may contain errors; check the original before relying on it.