The Verge’s investigation, citing internal ByteDance documents, revealed that a project known as Project Seed extensively employed the OpenAI API throughout various stages of LLM development, including model training and assessment. This revelation raised concerns regarding ByteDance’s compliance with OpenAI’s usage policies.
ByteDance’s Response and OpenAI’s Action
In response to the publication by The Verge, ByteDance issued a statement to the South China Morning Post. The company admitted to utilizing the OpenAI API in an experimental model for LLM exploration earlier in the year. However, ByteDance clarified that the project was not intended for public release and was terminated in April following the introduction of regulations by OpenAI, prohibiting the development of competing models using their products.
Consequently, OpenAI suspended ByteDance’s access to its service pending an investigation. The AI giant emphasized that even minimal usage that contradicts their terms and conditions could lead to corrective actions or account closure for the violating party.
Rise of LLM Development in China
The controversy comes amid a surge in large language model (LLM) development within China, notes NIX Solutions. As of July this year, over 130 LLMs had been developed by local firms and research institutes. This proliferation prompted Baidu founder and CEO Robin Li to remark on the excessive creation of basic AI models, labeling it as a “huge waste of resources.”