• The US accused Chinese AI firms of maliciously copying American technology.
• Techniques like distillation are alleged to be used for industrial-scale theft.
• Major companies named include DeepSeek, Alibaba, and Moonshot AI.
• The accusations complicate US-China relations before a key diplomatic meeting.
• Potential sanctions against Chinese firms are under consideration by US authorities.

The US government has officially accused several Chinese artificial intelligence companies, namely DeepSeek, Alibaba, and Moonshot AI, of maliciously copying technology from American firms. This has raised tensions ahead of a significant meeting between the leaders of the US and China, scheduled for later this month. According to US intelligence and law enforcement sources, these companies are engaging in practices that may involve government awareness to expedite their AI product development through a method called distillation, which allows the training of smaller AI models using data from larger ones. This method not only diminishes research costs but is perceived as a threat to national security, according to multiple sources including South China Morning Post, Reuters, and India Times.

Officials highlighted that the allegations stem from a joint advisory involving key US security agencies, which states that these AI firms are not only targeting proprietary functionalities of advanced American AI models but also tapping into a network of intermediaries to conduct their operations covertly. As described by the advisory, these efforts signify an industrial-scale campaign that undermines the integrity of American technological advancements and could enhance China's military and cyber capacities, as per reports from Reuters and India Times.

In light of these allegations, US Treasury Secretary Scott Bessent indicated that sanctions and further measures could be considered against any Chinese firm found to have engaged in intellectual property theft through such distillation activities. Notably, reports from companies like Anthropic elaborate that specific Chinese laboratories have been identified as having run extensive "high-volume knowledge distillation campaigns," further deepening the concerns about security and technology asset breaches, as reported by South China Morning Post, Reuters, and India Times.