Watch Desk posted an update
A Reuters review of more than 300 Chinese documents says the use of leading US AI models to train domestic rivals has spread from academic experiments into widespread industrial practice. Reuters also says Beijing is encouraging the approach.
Why it mattersThat points to model distillation becoming part of China’s AI industry, not just a research technique. The report’s account raises questions about how American model developers can protect their work, and whether export controls can contain know-how that travels through model outputs.
Discuss: If model outputs can help train competing systems, should AI developers restrict access to their strongest models, or would that do more harm than good?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.