OpenAI has alerted United States lawmakers to what it described as attempts by Chinese artificial intelligence startup DeepSeek to replicate advanced AI models developed by American companies.
In a memo sent to the House Select Committee on Strategic Competition, the firm accused the rival group of attempting to copy systems through technical workarounds and data extraction techniques.
Claims Of Distillation And Access Circumvention
The company said DeepSeek pursued “ongoing efforts to free-ride on the capabilities developed by OpenAI and other U.S. frontier labs.”
According to the document, employees allegedly bypassed access restrictions using masked routing methods to obtain model outputs.
“We have observed accounts associated with DeepSeek employees developing methods to circumvent OpenAI’s access restrictions and access models through obfuscated third-party routers and other ways that mask their source.”
“We also know that DeepSeek employees developed code to access U.S. AI models and obtain outputs for distillation in programmatic ways,” the memo added.
Distillation refers to using a more advanced AI system to evaluate responses from a newer model, transferring knowledge and improving performance without direct training data access.
Growing AI Competition Concerns
DeepSeek gained International attention after releasing models that rivaled leading Western systems, prompting policy concerns about technological competition between the United States and China.
OpenAI further warned that Chinese large language models were “actively cutting corners when it comes to safely training and deploying new models.”
The company stated it actively removes accounts suspected of using its services to train competing artificial intelligence products.
DeepSeek and parent company High-Flyer did not immediately comment on the allegations, leaving the dispute unresolved as regulators review the claims.