OpenAI accuses DeepSeek of distilling US AI models illegally
Synopsis
OpenAI has accused Chinese AI startup DeepSeek of using outputs from US artificial intelligence systems to train its own models. The company said the activity involved distillation, a method that uses responses from advanced AI systems as training material. OpenAI is reviewing potential policy violations and monitoring misuse of its technology. The allegation highlights growing competition between US and Chinese AI developers and raises concerns about intellectual property protection, access to advanced systems, and enforcement of artificial intelligence usage rules globally.
OpenAI accuses DeepSeek of distilling US AI models illegally, alleging the Chinese startup used outputs from American artificial intelligence systems to train its own chatbot. The dispute highlights growing concerns over intellectual property protection and competition between AI developers in the United States and China.
Key Highlights
- DeepSeek has not publicly responded in detail to reported allegations
- OpenAI alleges DeepSeek used US AI outputs to train competing artificial intelligence models
- Claims involve distillation, a technique using AI-generated responses as training data
- Internal memo cited evidence of China-based efforts to replicate US AI capabilities
OpenAI has alleged that Chinese artificial intelligence firm DeepSeek has been using the output of US-made artificial intelligence systems to train its models. Bloomberg News initially reported it and states that DeepSeek employed a technology called distillation to build rival chatbot technology.
Distillation enables developers to train a smaller model on the output of a larger AI model, allowing them to create new models with fewer computing needs, much faster. This would be able to create new models at a higher rate within less computing space. OpenAI claimed that there is evidence that groups in China used this method to imitate the capabilities of the advanced US systems.
The internal memo quoted in the reports indicated that OpenAI itself knows that it is being used to assist in developing rival AI tools using the output of its models. The firm did not post any technical information about the purported act but indicated that it is investigating the case and evaluating any possible contravention of its policies.
The AI firm has not publicly responded to the allegations.
OpenAI exposes its AI systems to application programming interfaces or API into software so developers can add AI functionality. The policies of the company have the limitation of the use of outputs, such as the limitation of using responses to train rival models.
The controversy indicates larger conflicts regarding the availability of high-tech technology, especially as states and corporations are trying to assert more control over intellectual property and keep AI creation within reach. It further brings up the rising issues surrounding AI systems training and the enforcement of usage policies by businesses.
OpenAI claimed it is still tracking how its services have been used and scrutinising possible misuse. The firm has not given information on whether it intends to sue the other party.
The accusation is made as the world competition between US and Chinese technology developers of artificial intelligence increases. Advanced AI systems require substantial computing resources and large datasets to train and are, therefore, useful when an existing model output is desired by a company to achieve faster development.
Follow Inspirepreneur Magazine for the latest business news.
Pooja Malik is a business journalist with over six years of experience covering startups, entrepreneurship, and emerging trends. She has previously worked with leading media platforms such as YourStory Media and BW BusinessWorld, where she reported on business, policy, and market developments. Currently, she serves as Editor at The Inspirepreneur Magazine, where she writes and edits stories across business, lifestyle, and travel, with a focus on clarity, accuracy, and reader relevance.
You Might Also Like
Global Military Spending Rises in 2025 Despite US Decline, Says SIPRI
5 Surprising Ways to Turn Conversations into Business Collaborations
