OpenAI Accuses DeepSeek of Using Its AI Technology

OpenAI has formally written to U.S. lawmakers alleging that Chinese artificial intelligence company DeepSeek is leveraging American AI technology to develop its own competing systems.

In the letter, OpenAI claims that DeepSeek is benefiting from the extensive research and computational resources invested by U.S.-based AI laboratories in order to strengthen its position in the rapidly evolving artificial intelligence sector.

Use of “Distillation” Technique

According to OpenAI, DeepSeek employs a technique known as “distillation.” This method allows a smaller or newer AI model to be trained by learning from the outputs generated by more advanced systems such as ChatGPT and other leading models.

Through distillation, a smaller model can achieve high performance in a relatively short time without requiring the same level of funding, data, or computational power that would be needed to train a model from scratch.

OpenAI argues that this approach enables DeepSeek to compete globally while significantly reducing development costs.

Security and Safeguard Concerns

OpenAI also stated that it has detected attempts by individuals allegedly linked to DeepSeek to bypass its technological safeguards. The company claims that sophisticated technical methods were used to collect data and insights from American AI systems.

Furthermore, OpenAI accused certain Chinese firms of overlooking key safety protocols during the development and deployment of new AI systems in order to accelerate their release to market.

DeepSeek’s Rapid Rise

DeepSeek, developed by Chinese firm High-Flyer, has gained international attention for demonstrating performance comparable to some leading American AI models — reportedly at a significantly lower training cost.

The dispute highlights growing tensions between the United States and China in the race for AI leadership, as concerns over intellectual property, national security, and technological dominance continue to intensify.

The outcome of this disagreement could have broader implications for global AI governance and cross-border technology collaboration.