1 week ago
2
ARTICLE AD BOX
DeepSeek is under scrutiny for alleged OpenAI data breach. (Photo: Shutterstock)
Microsoft and OpenAI are examining whether a group associated with DeepSeek gained unauthorised access to data from the ChatGPT maker’s technology. According to Bloomberg News, which broke the story, the investigation follows security concerns raised by Microsoft, which detected unusual data extraction activity linked to OpenAI’s systems.
According to the report, Microsoft’s security researchers observed individuals they believe are connected to DeepSeek transferring significant volumes of data through OpenAI’s application programming interface (API). The API serves as a central gateway for developers and enterprises to integrate OpenAI’s technology into their applications. As OpenAI’s largest investor, Microsoft flagged the activity, prompting an internal review.
DeepSeek, a Chinese artificial intelligence (AI) startup, has exploded in popularity in recent days as an alternative to more established US competitors ChatGPT and Claude. The company’s AI assistant has overtaken the former in downloads on Apple’s App Store in the US, contributing to a general decline in technology stock prices on Monday.
The situation has also attracted the attention of US officials, including David Sacks, a senior White House adviser on AI and cryptocurrency. Speaking to Fox News, Sacks stated that there was “substantial evidence” that DeepSeek had extracted knowledge from OpenAI’s models.
An OpenAI spokesperson, without directly naming DeepSeek, noted ongoing attempts by foreign companies to replicate leading US AI models. The spokesperson stated that OpenAI employs protective measures, including a controlled approach to releasing advanced capabilities, and emphasised the importance of working with the US government to safeguard AI technology.
The White House is conducting a national security assessment of DeepSeek’s AI technology, press secretary Karoline Leavitt confirmed on Tuesday. The National Security Council is leading the evaluation, reviewing potential risks posed by foreign AI developments.
Leavitt described the situation as a signal for the US AI industry to prioritise domestic innovation and technological leadership. The previous Biden administration imposed extensive export restrictions on AI chips and semiconductor manufacturing equipment, aiming to slow China’s progress in AI.
President Donald Trump addressed the issue on Monday, stating that the rise of the Chinese AI app should push US companies to enhance their competitiveness. “The release of DeepSeek AI from a Chinese company should be a wake-up call for our industries that we need to be laser-focused on competing to win,” Trump said. He also noted that China’s development of a cheaper and faster AI model highlighted the need for the US to remain competitive in the sector.
Italian data regulator investigates DeepSeek’s information practices
In a separate development, Italy’s data protection authority, the Guarantor for the Protection of Personal Data, has requested information from Hangzhou DeepSeek Artificial Intelligence and Beijing DeepSeek Artificial Intelligence regarding their data handling practices.
The regulator has asked the companies to disclose what personal data they collect, the sources of this information, and the legal basis for its processing. Authorities have also inquired whether the data is stored on servers in China. Additional requests include clarification on the training methods used for DeepSeek’s AI models. If the company employs web scraping, the regulator has asked how both registered and unregistered users have been informed about data collection.
DeepSeek has been given 20 days to respond to the Italian regulator’s inquiries. Fears about misuse of user input data by the firm are legitimate and employers should educate staff on the ramifications of using its models, argued analysts from Forrester.
“The massive downloads of DeepSeek mean that thousands (and even millions of users) are experimenting and uploading what could be sensitive information into the app,” they wrote. “According to its privacy policy, DeepSeek explicitly says it can collect “your text or audio input, prompt, uploaded files, feedback, chat history, or other content” and use it for training purposes. It also states it can share this information with law enforcement agencies [and] public authorities at its discretion.”
