China’s JoyIn claims similarities between its model and OpenAI developments
JoyIn, a Chinese humanoid robotics startup from Suzhou, has claimed similarities between the technical approaches and webpage designs of its Aether model and OpenAI’s recent materials. JoyIn CEO Guo Zhenjie also said the company has begun the process of filing a lawsuit.
JoyIn’s claims of similarities
According to CNBC, in an open letter in Chinese, Guo said that several weeks earlier the startup had presented the concept of an “alien visitor” AI model in Silicon Valley. After that, on September 6, OpenAI’s chief scientist published a piece titled “An Alien Mind.”
Guo questioned the similarities of approaches such as recursive self-improvement and the use of artificial intelligence to optimize computing capacity. He also pointed to the space-themed design of OpenAI’s GPT-6 Astra webpage and the Aether website. According to the JoyIn chief, the Aether website was launched two months earlier.
CNBC notes that it was unable to independently verify these claims. The outlet also states that some of the mentioned concepts are already used in artificial intelligence research, although companies may combine and implement them differently. OpenAI did not immediately comment on CNBC’s request.
More current news is available on the UA.News Telegram channel Telegram.
Aether model for robots
On September 10, JoyIn publicly announced the Aether model for controlling humanoid robots. The company claims that its perception-based, rather than text-based, approach enables robots to complete tasks with a 90% success rate on the first attempt.
Development head Zhu Mingxuan previously worked at the US company Figure on models that helped humanoid robots reproduce human movements. According to her, Aether reduced training time by two-thirds. She also announced plans to open-source parts of the model related to touch, but not components concerning energy use.
Previously, US-based Anthropic repeatedly alleged unauthorized “distillation” of its models by Chinese companies to develop their own AI systems. On Tuesday, a US cybersecurity agency reported that six Chinese companies, including DeepSeek and Alibaba, had distilled models from Anthropic, Google and OpenAI.
