OpenAI Alleges DeepSeek Used Proprietary Models Without Permission As Microsoft Investigates AI Security Risks: 'There's Substantial Evidence'

In This Article:

OpenAI Alleges DeepSeek Used Proprietary Models Without Permission As Microsoft Investigates AI Security Risks: 'There's Substantial Evidence'
OpenAI Alleges DeepSeek Used Proprietary Models Without Permission As Microsoft Investigates AI Security Risks: 'There's Substantial Evidence'

OpenAI announced it has uncovered evidence that Chinese artificial intelligence startup DeepSeek allegedly used its proprietary models for unauthorized training purposes, the ChatGPT maker told the Financial Times, intensifying concerns over intellectual property protection in the AI sector.

What Happened: The San Francisco-based company, partnering with Microsoft Corp. (NASDAQ:MSFT), discovered suspicious activity last fall when security researchers identified individuals reportedly linked to DeepSeek extracting substantial data through OpenAI’s API, Bloomberg reported, citing sources. The investigation led to the blocking of suspected accounts for violating terms of service.

Don't Miss:

This development emerges as OpenAI navigates complex discussions about its transition to a for-profit entity, with its board working to value Microsoft’s stake while negotiating a $30 billion valuation for its new charitable division.

The restructuring talks, ongoing since September, aim to split OpenAI into two entities, with the charitable arm maintaining a stake in the newly formed public benefit corporation.

OpenAI, Microsoft, and DeepSeek did not immediately respond to Benzinga's request for comment.

David Sacks, Trump’s AI and crypto tsar in an interview with Fox on Tuesday said, “There's substantial evidence” that Deepseek “distilled knowledge out of OpenAI's models”. Distillation is the process where smaller models obtain better performance by using outputs from larger models.

Trending: Nancy Pelosi Invested $5 Million In An AI Company Last Year — Here's How You Can Invest In Multiple Pre-IPO AI Startups With Just $1,000

Adding to the momentum, OpenAI CEO Sam Altman hinted at expanding collaboration with Microsoft on Wednesday, posting on X: “next phase of the msft x oai partnership is gonna be much better than anyone is ready for!!” alongside a photo with Microsoft CEO Satya Nadella.


See Also: If there was a new fund backed by Jeff Bezos offering a 7-9% target yield with monthly dividends would you invest in it?

Why It Matters: The launch of DeepSeek sent ripples through tech markets, initially causing volatility among AI-related stocks. NVIDIA Corp. (NASDAQ:NVDA) experienced particular turbulence, dropping 17% on Monday before rebounding 9% on Tuesday, as investors processed DeepSeek’s claims of achieving comparable AI performance at significantly lower costs.