Emre Yumlu 
Security researchers at Microsoft, OpenAI’s largest investor, detected suspicious data extraction from the company’s API in the fall of 2024.
Microsoft and OpenAI Investigate DeepSeek-Linked Data Leak
Microsoft and OpenAI are investigating whether a group connected to the Chinese AI startup DeepSeek extracted data from OpenAI’s technology infrastructure without permission.
According to Bloomberg’s report on January 29, Microsoft’s security researchers detected large-scale suspicious data extraction through OpenAI’s API in late 2024. As OpenAI’s largest investor, Microsoft reported the issue to the company, noting that it could be a violation of ChatGPT’s terms of service or a breach of data collection restrictions.
On January 20, DeepSeek, based in China, unveiled a new AI model called R-1. Despite significantly lower production costs, the model is said to be competitive with ChatGPT. Following the announcement, U.S. tech and AI stocks saw billions of dollars in value losses.
David Sacks, White House Crypto and AI Advisor, stated in an interview with Fox News on January 28 that there is evidence DeepSeek used OpenAI’s model outputs to train its own AI. Sacks explained that this method, known as “distillation,” allows one AI model to develop similar capabilities by training on the outputs of another model.
Sacks said, “There is significant evidence that what DeepSeek is doing is distilling information from OpenAI’s models, and I don’t think OpenAI is very happy about this.”
OpenAI and U.S. Navy Take Action Against DeepSeek
While OpenAI expressed concerns about Chinese companies attempting to copy U.S. AI models, it refrained from commenting specifically on DeepSeek.
An OpenAI spokesperson told Bloomberg, “Companies in the People’s Republic of China and others are attempting to distill the leading AI models from the U.S.” However, the company did not make a specific statement about DeepSeek’s activities.
The investigation conducted by OpenAI and Microsoft is looking into whether the company’s terms of service were violated and whether data access restrictions were bypassed. This situation highlights the intensifying AI technology competition between the U.S. and China.
Amid these developments, CNBC reported on January 28 that the U.S. Navy has completely banned the use of DeepSeek. This decision was made due to concerns that the Chinese government might use AI to gain access to sensitive data. The decision was announced in an email sent to personnel on January 24.
The message emphasized that DeepSeek posed “security and ethical risks” and that Navy members should “not use this model under any circumstances.”
Also, you can freely share your thoughts and comments about the topic in the comment section. Additionally, please follow us on our Telegram, YouTube and Twitter channels for the latest news and updates.