OpenAI has claimed that DeepSeek, a new AI chatbot developed in China, might have trained its own open-source model using OpenAI’s proprietary technology. In other words, they’re alleging that the the artificial intelligence start-up didn’t build everything from scratch but instead relied on OpenAI’s models to develop its competitor.
The Financial Times cited the San Francisco-based ChatGPT maker as having seen evidence of “distillation,” the process of transferring knowledge from a large model to a smaller one. As such, this would be seen as a breach of OpenAI’s terms of service.
Knowledge distillation has been put to good use in all sorts of areas, from natural language processing and speech recognition to image recognition and object detection. But in recent years, it’s become especially important for large language models (LLMs). When it comes to LLMs, distillation has proven to be an important way to pass on advanced skills from top-tier proprietary models to smaller, more accessible open-source ones. In this case, DeepSeek has been accused of breaching intellectual property.
A source told the FT: “The issue is when you [take it out of the platform and] are doing it to create your own model for your own purposes.”
👀 DAVID SACKS: “There is substantial evidence that what DeepSeek did here is they distilled the knowledge out of OpenAI’s models, and I don’t think OpenAI is very happy about this.” pic.twitter.com/IYXKwBbUh1
— Chief Nerd (@TheChiefNerd) January 29, 2025
US President Donald Trump’s new AI and cryptocurrency ‘tsar,’ David Sacks, also told Fox News: “There’s substantial evidence that what DeepSeek did here is they distilled knowledge out of OpenAI models, and I don’t think OpenAI is very happy about this.
“I think one of the things you’re going to see over the next few months is our leading AI companies taking steps to try to prevent distillation.”
Still, OpenAI CEO Sam Altman did publicly applaud the open-source R1 model, calling it “impressive.”
DeepSeek is making investors nervous
DeepSeek’s release of its R1 model has sent shockwaves through global financial markets. The Chinese company seems to have pulled off results comparable to its biggest rivals, despite using far less money and computing power.
This has made investors rethink a key assumption behind the recent AI-driven stock market boom. The biggest AI players, or “hyperscalers,” need massive amounts of computing power to stay ahead. The uncertainty hit Nvidia particularly hard, causing its stock to suffer the largest single-day drop in market value in history on Monday (Jan. 27), though it managed to regain some ground on Tuesday.
OpenAI isn’t exactly free from criticism when it comes to intellectual property. The company is currently dealing with a lawsuit, led by The New York Times, where media companies accuse it of using their data without permission.
Still, these latest allegations against DeepSeek could add another layer to the ongoing tech battle between the US and China, making things even more complicated.
ReadWrite has reached out to OpenAI for comment.
Featured image: Canva / DeepSeek / ReadWrite