White House AI czar David Sacks has raised concerns that Chinese AI startup DeepSeek may have used OpenAI’s technology to develop its groundbreaking R1 model, potentially engaging in intellectual property theft. Speaking to Fox News on Tuesday, Sacks highlighted a controversial AI training technique called distillation, where a smaller “student model” learns from a larger, more complex “teacher model” by mimicking its reasoning processes and extracting knowledge.
DeepSeek’s R1 model has sent shockwaves through Silicon Valley by matching the capabilities of leading AI models while using significantly less computational power — the expensive infrastructure that typically drives AI development costs into the hundreds of millions of dollars. This achievement sparked a stock market panic earlier this week, wiping hundreds of billions of dollars in market value from AI chip manufacturers like Nvidia, as investors questioned whether massive computing investments would remain necessary for competitive AI development.
“There’s substantial evidence that what DeepSeek did here is they distilled the knowledge out of OpenAI’s models,” Sacks stated, though he did not provide specific evidence to support the claim. He added that “I don’t think OpenAI is very happy about this” and suggested that companies would implement measures to prevent such distillation practices, which “would definitely slow down some of these copycat models.”
The situation presents a complex ethical and legal landscape. OpenAI itself offers model distillation services through its API, allowing smaller models to create datasets for training purposes. However, the company’s terms of service explicitly prohibit users from “copying” its services or using its models to build competing products. OpenAI has faced its own scrutiny over training practices, currently defending against multiple copyright infringement lawsuits from media organizations including The New York Times.
Neither Microsoft (OpenAI’s primary investor and partner), OpenAI, nor DeepSeek responded to requests for comment. DeepSeek’s silence may be partly attributed to the timing, as the request was made during the Lunar New Year holiday. The allegations raise significant questions about international AI competition, intellectual property protection, and the future of AI development as Chinese companies challenge American technological dominance.
Key Quotes
They can essentially mimic the reasoning process that they learn from the parent model and they can suck the knowledge out of the parent model
David Sacks, White House AI and crypto czar, explained how the distillation technique works, describing how smaller AI models can extract knowledge from larger, more sophisticated models to replicate their capabilities.
There’s substantial evidence that what DeepSeek did here is they distilled the knowledge out of OpenAI’s models. I don’t think OpenAI is very happy about this
Sacks made this accusation on Fox News, though he did not provide specific evidence. This statement represents the most direct allegation from a senior U.S. government official regarding potential intellectual property theft in AI development.
That would definitely slow down some of these copycat models
Sacks suggested that AI companies would implement protective measures against distillation, indicating that the industry may move toward more restrictive access to prevent competitors from replicating their models.
Our Take
This controversy exposes a fundamental vulnerability in the current AI development ecosystem. The irony is striking: OpenAI, which itself faces copyright infringement lawsuits for training on protected content, now finds its own models potentially being used as training data. This highlights how distillation exists in a legal gray area — it’s a legitimate technique that OpenAI itself offers commercially, yet can be considered theft when used to build competing products. The market’s violent reaction, erasing hundreds of billions in value, suggests investors are grappling with a paradigm shift: if advanced AI can be developed cheaply through distillation, the moat around expensive computational infrastructure may be narrower than believed. This could accelerate calls for stronger technical protections and clearer international agreements on AI development practices, particularly as geopolitical tensions between the U.S. and China intensify around technological supremacy.
Why This Matters
This controversy represents a pivotal moment in the global AI race, particularly in U.S.-China technological competition. If DeepSeek successfully created a competitive AI model at a fraction of the typical cost, it challenges the prevailing assumption that AI leadership requires massive capital investments in computing infrastructure. This has immediate implications for AI chip manufacturers and cloud computing providers whose business models depend on escalating computational demands.
The distillation allegations also highlight critical gaps in AI intellectual property protection. Unlike traditional software, AI models can potentially be reverse-engineered through their outputs, making enforcement of usage terms extremely difficult. This could force the industry to develop new technical safeguards and legal frameworks.
For businesses and policymakers, this situation underscores the tension between open AI development and competitive advantage. Companies must balance making their AI accessible through APIs (generating revenue and adoption) against the risk of enabling competitors to replicate their technology. The outcome of this controversy could reshape how AI companies approach model deployment, API access, and international collaboration in an increasingly fragmented technological landscape.
Recommended Reading
For those interested in learning more about artificial intelligence, machine learning, and effective AI communication, here are some excellent resources:
Recommended Reading
Related Stories
- OpenAI CEO Sam Altman Hints at Potential Restructuring in 2024
- OpenAI’s Valuation Soars as AI Race Heats Up
- OpenAI CEO Sam Altman’s Predictions on How AI Could Change the World by 2025
- Elon Musk Drops Lawsuit Against ChatGPT Maker OpenAI, No Explanation
- The Artificial Intelligence Race: Rivalry Bathing the World in Data