OpenAI has raised concerns to the US government regarding Chinese AI startup DeepSeek, which allegedly aims to replicate American AI technologies by bypassing platform safeguards. In a memo to the US House Select Committee on Strategic Competition, OpenAI accused DeepSeek of “free-riding” on advancements developed by US labs through methods termed “distillation.” The memo outlines how DeepSeek employees reportedly developed techniques to access and extract model outputs from OpenAI’s systems using obfuscated third-party methods. OpenAI emphasized that while distillation has legitimate applications, unauthorized use can lead to unregulated AI systems lacking consumer safety measures. Additionally, the report accused DeepSeek of exhibiting strong pro-Communist Party biases in its model responses, often censoring sensitive topics related to the CCP. OpenAI underscored that the US must maintain an advantage in AI technology and compute power to support further advancements amid growing competition with China.
Source link
