OpenAI says DeepSeek may have 'inapproriately' used its data

[duluxe]

https://indianexpress.com/article/explained/explained-sci-tech/deepseek-openai-technology-9807132/


OpenAI says DeepSeek, its sudden Chinese rival, may have "inappropriately” taken data from its model to spin up its own artificial intelligence chatbot.

DeepSeek released a surprisingly effective and inexpensive Large Language Model, or LLM, on Monday, shocking U.S. markets and causing the stock of the top U.S. chip manufacturer, Nvidia, to tumble.


The Financial Times first reported OpenAI’s claim.

An OpenAI spokesperson said in an emailed statement to NBC News Wednesday that unnamed Chinese companies are actively working to catch up to American AI companies through a tactic known as distillation, where an LLM is trained using data generated by another LLM, and said it believes DeepSeek may be one of those companies.

“We are aware of and reviewing indications that DeepSeek may have inappropriately distilled our models, and will share information as we know more,” the spokesperson said, adding that the company was not accusing DeepSeek of a security breach.

Distillation is often prohibited in LLMs’ terms of service, but is common in the industry.

DeepSeek did not respond to a request for comment.

OpenAI has itself been accused of building ChatGPT by inappropriately accessing content it did not have the rights to. It is the subject of multiple lawsuits along those lines, including from The New York Times, which claims the company built its model in part by ingesting millions of the newspaper’s stories without permission.

On Monday, OpenAI CEO Sam Altman posted a congratulations to DeepSeek on X, saying it had developed “an impressive model, particularly around what they’re able to deliver for the price.”

The OpenAI spokesperson indicated the White House may get involved in the dispute.

“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 said.

On Tuesday, President Donald Trump’s AI adviser, David Sacks, told Fox News that he believed DeepSeek did distill from OpenAI, but stopped short of declaring that the White House would take action.

“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 and prevent distillation,” he said. “That would definitely slow down some of these copycat models.”

Brian Hughes, a spokesperson for the National Security Council, said in an email to NBC News that the White House was looking into the broader issue but didn't specifically address OpenAI's distillation accusation.

“At the President’s direction, the NSC and others in the U.S. government are working in many ways to address concerns involving AI, China and data security. As the President has underscored, U.S. policy is to ensure that the United States leads the world in AI.”