ByteDance’s accounts have been banned by OpenAI after reports surfaced that ByteDance was using OpenAI’s output to train their own AI models. Alex Heath from The Verge reported that ByteDance was using their own version of LLM called Project Seed to train their models using OpenAI’s output, even though they have stopped using the output for direct model training purposes. However, they still use it to measure the quality of their own models.
OpenAI has confirmed the temporary ban on ByteDance’s accounts while they investigate the matter. It is unclear whether Microsoft Azure, through which ByteDance accessed OpenAI’s API, has also banned their accounts.
ByteDance has denied these reports, stating that they use OpenAI’s API through Azure to serve their customers outside of China, in addition to using their own Duobao in China.
Most LLM AI service providers have strict usage restrictions, prohibiting their use for competitive development.
TLDR: OpenAI temporarily bans ByteDance’s accounts after discovering that they were using OpenAI’s output to train their own AI models. ByteDance accesses OpenAI’s API through Microsoft Azure, and it is unclear if Microsoft has also banned their accounts. ByteDance denies the allegations and states that they use OpenAI’s API to serve customers outside of China, alongside their own Duobao service in China. Most LLM AI service providers have restrictions against competitive development.
Leave a Comment