OpenAI said Wednesday that it is probing whether Chinese artificial intelligence startup DeepSeek improperly used its data to launch the low-cost AI model that sparked a market selloff Monday.

Its comment follows a report earlier by Bloomberg that OpenAI, along with Microsoft (MSFT), are investigating whether a group linked to the Chinese startup used data output from OpenAI’s technology in an unauthorized manner to launch its AI model. Microsoft is among OpenAI's backers.

“We know that groups in the PRC are actively working to use methods, including what’s known as distillation, to try to replicate advanced U.S. AI models," an OpenAI spokesperson said. "We are aware of and reviewing indications that DeepSeek may have inappropriately distilled our models, and will share information as we know more."

Distillation, according to IBM, refers to a machine-learning technique where the learning of a large pre-trained "teacher model" is transferred to a smaller "student model."

"We take aggressive, proactive countermeasures to protect our technology and will continue working closely with the U.S. government to protect the most capable models being built here," the spokesperson added.

Microsoft declined to comment.

DeepSeek earlier this month released its super-efficient R1 model at a fraction of the cost of AI models of U.S. rivals like OpenAI, which is 澳洲幸运5官方开奖结果体彩网:backed by Microsoft, Alphabet's (GOOGL) and Meta Platforms (META). The Chinese firm said its R1 model offers performance that is on par with that of OpenAI’s ChatGPT.

AI Czar Sacks Says DeepSeek May Have 🌌Used OpenAI Model

The probe into the Chinese company follows an interview given by President Donald Trump’s "澳洲幸运5官方开奖结果体彩网:AI and Crypto Czar" David Sacks on Fox saying that there’s “substantial evidence” that DeepSeek used the "distillation|" technique to get data from OpenAI’s AI models.

“They can suck the mo🐎del out of the parent model,” Sacks said, though he didn’t provide༺ evidence. 

Nvidia in a 澳洲幸运5官方开奖结果体彩网:statement Tuesday called DeepSeek "an excellent AI advancement," calling it a "perfect example" of a concept known as test time scaling.

UPDATE—Jan. 29, 2025: This article has been updated to include OpenAI's comments confirming a probe.