Microsoft investigating if DeepSeek illegally used OpenAI’s data

Microsoft investigating if DeepSeek illegally used OpenAI’s data
HIGHLIGHTS

Microsoft and OpenAI are looking into whether data from OpenAI’s technology was obtained unlawfully by DeepSeek.

Microsoft discovered suspicious activity last fall, when they noticed individuals connected to DeepSeek exfiltrating large amounts of data using the OpenAI API.

This API allows software developers to pay for a license to integrate OpenAI’s powerful AI models into their own applications.

Microsoft and OpenAI are looking into whether data from OpenAI’s technology was obtained unlawfully by DeepSeek, a Chinese artificial intelligence startup. Microsoft’s security researchers discovered suspicious activity last fall, when they noticed individuals connected to DeepSeek exfiltrating large amounts of data using the OpenAI application programming interface (API), reports Bloomberg, citing people familiar with the matter. This API allows software developers to pay for a license to integrate OpenAI’s powerful AI models into their own applications.

After noticing the suspicious activity, Microsoft, OpenAI’s biggest investor and technology partner, alerted OpenAI. If the activity is confirmed, it could be a breach of OpenAI’s terms of service. It could also suggest that DeepSeek may have found a way to bypass restrictions on how much data they could obtain.

Also read: Who Is Liang Wenfeng? The DeepSeek founder shaking up the AI world

Earlier this month, DeepSeek launched a new open-source AI model called R1, which is designed to mimic human reasoning. R1 is said to rival or even outperform top AI models from leading US companies like OpenAI, Google, and Meta Platforms.

DeepSeek claims R1 can excel in various industry benchmarks, including mathematical tasks and general knowledge, all while being developed at a fraction of the cost of American models. This has raised concerns that DeepSeek’s new model could threaten the dominance of US-based AI companies, which contributed to a significant drop in the market value of AI-related stocks, including Microsoft, Nvidia, Oracle, and Google’s parent company, Alphabet, according to the report.

Also read: Is Deepseek sending data to China? Here is what it’s privacy policy says

David Sacks, former President Donald Trump’s AI czar, stated there is “substantial evidence” that DeepSeek used OpenAI’s models to help develop their own technology. He explained that “what DeepSeek did here is they distilled knowledge out of OpenAI models.”

OpenAI responded to Sacks’ remarks, acknowledging that companies, particularly from China, often attempt to “distill” models from leading US AI companies. While OpenAI did not address DeepSeek directly, they emphasised that they are actively protecting their intellectual property and working closely with the US government to safeguard their advanced AI models from competitors and adversaries.

Ayushi Jain

Ayushi Jain

Tech news writer by day, BGMI player by night. Combining my passion for tech and gaming to bring you the latest in both worlds. View Full Profile

Digit.in
Logo
Digit.in
Logo