US accuses China of “industrial-scale” AI theft. China says it’s “slander.” (arstechnica.com)

The US government is poised to take action against China for what it describes as “industrial-scale theft” of intellectual property related to artificial intelligence. This accusation follows reports from companies like OpenAI and Google, which claim that Chinese firms have been using a method called distillation to clone their AI models, undermining the competitive edge of American technology. For instance, OpenAI has alleged that the Chinese model DeepSeek was trained on outputs from its systems, and Google highlighted attempts to duplicate its Gemini chatbot by exploiting its model extensively. This issue is significant for the AI/ML community as it raises pressing concerns about the integrity and protection of intellectual property in rapidly evolving AI technologies. The White House Office of Science and Technology Policy has acknowledged that these distillation tactics could allow China to catch up swiftly in the AI race, with Chinese actors reportedly using thousands of proxy accounts and sophisticated methods to bypass detection and extract proprietary information. In response, the US may strengthen legal frameworks to address these threats more effectively, with Congress contemplating designating model extraction as industrial espionage and imposing stringent penalties on those engaging in such activities.