Howard Lutnick, the Commerce Secretary nominee who testified before the U.S. Senate Tuesday, claimed DeepSeek cheated when its revolutionary new AI model spanked top western AI companies this week.
“They stole things, they broke in, they’ve taken our IP,” Lutnick said, reflecting broader U.S. government concerns about Chinese companies potentially misappropriating U.S. technology.
The White House joined a growing backlash against DeepSeek on Wednesday, assessing potential national security risks tied to its AI. Critics, including Western firms, accuse the Chinese upstart of rule-breaking, espionage, and market manipulation.
It started late Monday after a gruesome day in tech stocks when OpenAI floated the idea that DeepSeek had used a technique called “distillation” (using a big model as an example to guide the answers provided by smaller models) to extract knowledge from its models—a violation of their terms of service.
David Sacks, Donald Trump’s AI adviser, piled on, adding that there was “substantial evidence” of DeepSeek “sucking knowledge” out of OpenAI’s technology.
The ChatGPT maker confirmed that they are investigating the issue.
“We are aware of and reviewing indications that DeepSeek may have inappropriately distilled our models, and will share information as we know more,” an OpenAI spokesperson told Axios.
The pushback came in response to DeepSeek’s impact on Nvidia’s market cap, erasing close to $600 billion—a record wipeout—on Monday after launching a new kind of AI model last week that supposedly matches U.S. capabilities at a mere fraction of the cost.
Hedge fund manager Bill Ackman—the financier who jokingly compared Trump to God—suggested that DeepSeek’s founders might have launched the model for free to short Nvidia stock. DeepSeek is backed by a quant hedge fund, after all.
“What are the chances that DeepSeek AI’s hedge fund affiliate made a fortune yesterday with short-dated puts on Nvidia, power companies, etc.? A fortune could have been made,” Ackman tweeted.
Though President Donald Trump initially called DeepSeek’s model a “wake-up call” for the industry and a “positive thing,” the U.S. Navy has already reached a different conclusion.
According to CNBC, the military branch ordered personnel late last week to steer clear of DeepSeek’s technology “in any capacity,” citing “potential security and ethical concerns associated with the model’s origin and usage.
In an email to the troops Friday, the Navy said it’s “imperative” for any team member to avoid DeepSeek “for any work-related tasks or personal use.”
Meanwhile, across the Atlantic, Italian regulators are also taking a hard look at DeepSeek.
The country’s Data Protection Authority gave DeepSeek 20 days to explain exactly what personal data they collect, where they store it, and how they use it to train their AI system—despite all of those answers being publicly available in DeepSeek AI’s Privacy Policy.
For its part, DeepSeek has thus far remained silent and has not replied to Decrypt (or anyone else) for comment.
“Karma’s a Bitch.”
You don’t need artificial intelligence to predict the Internet’s swift and merciless response to the whole affair:
“OpenAI is unhappy that DeepSeek trained on their data without consent or compensation,” tweeted Toby Walsh, Chief Scientist at the University of New South Wales’s AI Institute. “Oh, the irony for all us authors unhappy with OpenAI for training on our data without consent or compensation!” he posted on X.
“DeepSeek may well have broken OpenAI’s Terms of Service and distilled their intellectual property without permission. OpenAI may well have done analogous things to YouTube, New York Times, and countless artists and writers,” AI researcher and writer Gary Marcus tweeted. “Karma is a bitch.”
And in a lengthy rant on his blog, tech industry critic Edward Zitron said the controversy was symptomatic of deeper issues in American tech.
He argued it revealed fundamental problems with U.S. tech companies. “This isn’t about China — it’s so much fucking easier if we let it be about China,” Zitron fulminated. “It’s about how the American tech industry is incurious, lazy, entitled, directionless and irresponsible.”
Far calmer was Perplexity CEO Aravind Srinivas—who already adapted and integrated DeepSeek into his search engine.
Srinivas offered a technical defense of DeepSeek to allegations that it had copied OpenAI. “There’s a lot of misconception that China ‘just cloned’ the outputs of OpenAi,” he tweeted. “This is far from true and reflects an incomplete understanding of how these models are trained in the first place.
“DeepSeek R1 has figured out (Reinforcement Learning) fine-tuning… The main reason it’s so good is because it learned reasoning from scratch rather than imitating other humans or models.”
In other words, DeepSeek did the work.
Indeed, Zitron, clearly unnerved by the insinuations that DeepSeek did something dishonest, pointed to OpenAI and Anthropic as examples of what he called “the antithesis of Silicon Valley” and being more concerned with marketing than innovation.
His conclusion? The rush to frame DeepSeek as a Chinese threat conveniently masks a more uncomfortable truth: U.S. tech companies have lost their drive to create meaningful solutions.
“Personally, I genuinely want OpenAI to point a finger at DeepSeek and accuse it of IP theft, purely for the hypocrisy factor,” Zitron said. “This is a company that exists purely from the wholesale industrial larceny of content produced by individual creators and internet users, and now it’s worried about a rival pilfering its own goods? Cry more, Altman, you nasty little worm.”
Edited by Josh Quittner and Sebastian Sinclair
Generally Intelligent Newsletter
A weekly AI journey narrated by Gen, a generative AI model.
