Cryptopolitan
--
Microsoft Investigating DeepSeek-linked Group in OpenAI Data Scraping: Report
Microsoft has reportedly launched an investigation into Chinese AI startup, DeepSeek. The OpenAI sponsor alleges the firm used ChatGPT to build its competing R-1 model.
This investigation comes after researchers at Microsoft, one of OpenAI’s biggest investors, raised warnings about a group of users, believed to be linked to DeepSeek, pulling large amounts of data from OpenAI’s application programming interface (API), according to a Financial Times report
Microsoft researchers detect usage anomalies with OpenAI API
In the fall of 2024, Microsoft security researchers discovered unknown large-scale data harvests using OpenAI’s API. After the discovery, Microsoft promptly notified OpenAI of this suspicious activity, which was considered an illegal data scraping attempt and a breach of OpenAI’s terms of service.
China-based AI startup DeepSeek debuted its R-1 model earlier this week. This model is now at the core of Microsoft and OpenAI’s dispute. This comes after the application for the new AI model went viral online for its high-speed and accurate performance at a relatively lower cost than OpenAI’s ChatGPT. It also outperformed several other models belonging to US firms Meta and X in various benchmarks.
Deepseek disclosed that it only cost a few million dollars to train and run its new model due to the company’s decision to use cost-efficient chips. This news sent shockwaves through the US market. Nvidia, the leading producer of AI chips, saw its market cap plunge amid growing fears that the Chinese firm could challenge US companies’ dominance in the AI space.
Did DeepSeek train its AI with OpenAI models?
According to US President Donald Trump’s crypto and AI czar, David Sacks, there is compelling evidence that DeepSeek used OpenAI’s model outputs to train its own AI. This technical term for the process is called distillation, in which an AI model imitates another’s capabilities by learning from its outputs.
Sacks said, “There’s substantial evidence that what DeepSeek did here is they distilled knowledge out of OpenAI models, and I don’t think OpenAI is very happy about this.”
OpenAI stated that AI businesses in China and elsewhere are continuously attempting to take insights from US-developed models, but didn’t outright accuse DeepSeek of misconduct.
An OpenAI spokesperson told Bloomberg: “We know PRC-based companies — and others — are constantly trying to distill the models of leading US AI companies”.
As Microsoft’s investigation into DeepSeek proceeds, US officials are also taking prompt action to restrict concerns.
On January 24, the US Navy issued a directive that prohibits its personnel from engaging with DeepSeek in any capacity. CNBC reported that the prohibition is due to concerns that the Chinese government might use the AI model to digitally gain access to sensitive data.
Cryptopolitan Academy: FREE Web3 Resume Cheat Sheet - Download Now
This investigation comes after researchers at Microsoft, one of OpenAI’s biggest investors, raised warnings about a group of users, believed to be linked to DeepSeek, pulling large amounts of data from OpenAI’s application programming interface (API), according to a Financial Times report
Microsoft researchers detect usage anomalies with OpenAI API
In the fall of 2024, Microsoft security researchers discovered unknown large-scale data harvests using OpenAI’s API. After the discovery, Microsoft promptly notified OpenAI of this suspicious activity, which was considered an illegal data scraping attempt and a breach of OpenAI’s terms of service.
China-based AI startup DeepSeek debuted its R-1 model earlier this week. This model is now at the core of Microsoft and OpenAI’s dispute. This comes after the application for the new AI model went viral online for its high-speed and accurate performance at a relatively lower cost than OpenAI’s ChatGPT. It also outperformed several other models belonging to US firms Meta and X in various benchmarks.
Deepseek disclosed that it only cost a few million dollars to train and run its new model due to the company’s decision to use cost-efficient chips. This news sent shockwaves through the US market. Nvidia, the leading producer of AI chips, saw its market cap plunge amid growing fears that the Chinese firm could challenge US companies’ dominance in the AI space.
Did DeepSeek train its AI with OpenAI models?
According to US President Donald Trump’s crypto and AI czar, David Sacks, there is compelling evidence that DeepSeek used OpenAI’s model outputs to train its own AI. This technical term for the process is called distillation, in which an AI model imitates another’s capabilities by learning from its outputs.
Sacks said, “There’s substantial evidence that what DeepSeek did here is they distilled knowledge out of OpenAI models, and I don’t think OpenAI is very happy about this.”
OpenAI stated that AI businesses in China and elsewhere are continuously attempting to take insights from US-developed models, but didn’t outright accuse DeepSeek of misconduct.
An OpenAI spokesperson told Bloomberg: “We know PRC-based companies — and others — are constantly trying to distill the models of leading US AI companies”.
As Microsoft’s investigation into DeepSeek proceeds, US officials are also taking prompt action to restrict concerns.
On January 24, the US Navy issued a directive that prohibits its personnel from engaging with DeepSeek in any capacity. CNBC reported that the prohibition is due to concerns that the Chinese government might use the AI model to digitally gain access to sensitive data.
Cryptopolitan Academy: FREE Web3 Resume Cheat Sheet - Download Now
Disclaimer: Includes third-party opinions. No financial advice. May include sponsored content. See T&Cs.