According to The Verge, it was revealed that ByteDance had been secretly using OpenAI technology to develop its own large language model, which violated OpenAI's terms of service. At present, ByteDance's account has been suspended by OpenAI.
The Verge further stated that ByteDance's action was usually regarded as a "disrespectful" behavior, and it also directly violated OpenAI's terms of service. OpenAI's terms of service stipulate that the model exported by the company cannot be used to "develop any AI model that competes with our products and services". ByteDance purchased access to OpenAI through Microsoft, but Microsoft also made the same policy as OpenAI.
According to the revealed internal documents, ByteDance relies on OpenAI's application programming interface (API) at almost every development stage to develop its basic large language model, including training and evaluation model, which is code-named "Project Seed". Employees who participate in the "seed plan" are well aware of the adverse effects of this behavior. According to the chat records of ByteDance employees on the internal communication platform, the overseas version of Lark, they discussed how to whitewash the evidence through "data desensitization".