techend
By Tanveer Ahmed :
OpenAI has warned US lawmakers that Chinese artificial intelligence start-up DeepSeek is attempting to copy advanced American AI systems and use them to build its own competing models.
In a memo sent to the US House Select Committee on Strategic Competition, the ChatGPT developer alleged that DeepSeek is trying to “free-ride” on technology created by leading US AI companies. The document claimed that DeepSeek employees have been working to bypass OpenAI’s safeguards in order to gain indirect access to its models.
According to OpenAI, DeepSeek is using a technique known as “distillation,” where responses from powerful AI systems are analysed and reused to train newer models. This process, while common in research, becomes controversial when used without permission to replicate proprietary technology.
The memo further stated that accounts linked to DeepSeek staff were using disguised third-party systems to mask their identity and extract AI outputs programmatically. OpenAI said such activities violate its usage policies and undermine fair competition.
DeepSeek, which is based in Hangzhou, gained global attention last year after releasing advanced models that matched the performance of some leading US products. Its tools, including DeepSeek-V3 and DeepSeek-R1, have been praised by international developers and researchers.
However, OpenAI, led by CEO Sam Altman, raised concerns that some Chinese AI firms are prioritising speed and cost over safety and ethical deployment. The company added that it regularly blocks users suspected of trying to extract its model knowledge for commercial replication.
Neither DeepSeek nor its parent firm High-Flyer responded immediately to requests for comment.
The issue has intensified debate in Washington about technological security, intellectual property, and whether China could rapidly narrow the AI gap with the United States despite existing export controls and restrictions.
About The Author
