Chinese Humanoid Startup JoyIn Accuses OpenAI of Distilling Its Model, Plans Lawsuit

Guo said in the statement that "People often say major tech companies have intelligence networks monitoring the whole internet, this time I believe it, this is a direct distillation of us without any modifications." He said JoyIn had publicly presented its "extraterrestrial visitor" AI model framework in Silicon Valley a few weeks earlier, before OpenAI's chief scientist published "An Alien Mind" on Sept. 6. Guo questioned similarities in core technological approaches such as "recursive self-improvement" and the use of AI to optimize computing power.

Guo also pointed to similarities between the outer space-inspired design of OpenAI's GPT-6 Astra web page and the website for JoyIn's Aether model, which he claimed was released two months ago. He added that JoyIn has started the process of filing a lawsuit. CNBC was unable to independently verify Guo's claims. Some of the concepts he cited are existing parts of AI research more broadly, although companies connect and implement them differently. OpenAI did not immediately respond to a request for comment.

The dispute unfolded as U.S. companies and officials have raised concerns about model distillation. U.S.-based Anthropic has repeatedly flagged unauthorized "distillation" of its models by Chinese companies to improve their own AI capabilities. On Tuesday, a U.S. cybersecurity agency said six Chinese companies, including DeepSeek and Alibaba, distilled models from Anthropic, Google and OpenAI. JoyIn's Aether model, which Guo said he decided to publicly announce Thursday, claims its perceptive, rather than text-based, approach to robotic control enables humanoids to complete tasks with a 90% success rate on first attempt.

Zhu Mingxuan, who led the Aether model's development, said she left U.S. humanoid company Figure last year, where she had also worked on models for helping humanoid robots mimic human actions. She told CNBC earlier this week that Aether reduced training time by two-thirds, and that she planned to open-source parts of the model, such as those relating to touch, but not portions related to energy use.

Getting humanoids and AI models to perform as intelligently as humans across a variety of tasks has remained a challenge, amid a broader technology race between U.S. and Chinese companies. Garry Tan, chief executive of startup accelerator Y Combinator, told CNBC this month that he would "do nothing" about distillation and joined others in pointing out that U.S. companies have trained their AI models on data covered under copyright law. CNBC's Kate Rooney and Isabel O'Brien contributed to the report.

Editor's Summary JoyIn has accused OpenAI of directly distilling its model and says it has begun legal action, while releasing its Aether robotics model. OpenAI did not immediately comment, and CNBC could not verify the claims. The dispute adds to U.S. industry debate over distillation and copyright in AI training.