DeepSeek Just Surpassed ChatGPT, But Did It Cheat? Microsoft Thinks So

Microsoft and OpenAI are carrying out an investigation into whether a group linked to DeepSeek, a Chinese AI startup, accessed OpenAI’s technology without permission. The probe began after Microsoft’s security team flagged unusual data activity connected to OpenAI’s systems, according to a report by Bloomberg.

Researchers at Microsoft observed individuals, believed to be associated with DeepSeek, transferring large amounts of data through OpenAI’s application programming interface (API). This API is a key tool that allows developers and businesses to integrate OpenAI’s models into their own applications. As OpenAI’s biggest investor, Microsoft flagged the activity, triggering an internal review.

DeepSeek’s Rapid Rise Sparks Concerns

DeepSeek has quickly emerged as a competitor to OpenAI’s ChatGPT and Anthropic’s Claude. In a sign of its growing influence, DeepSeek’s AI assistant recently surpassed ChatGPT in downloads on Apple’s US App Store. This sudden success contributed to a dip in tech stock prices on Monday.

The situation has caught the attention of US officials. White House senior adviser David Sacks told Fox News there was “substantial evidence” that DeepSeek had extracted data from OpenAI’s models.

While OpenAI didn’t mention DeepSeek by name, a spokesperson acknowledged that foreign companies frequently try to replicate top US AI models. They added that OpenAI takes protective measures, including a controlled rollout of its most advanced features, and emphasised the importance of working with the US government to safeguard AI technology.

White House Launches National Security Review

The US government has now launched a national security review of DeepSeek’s AI technology. White House press secretary Karoline Leavitt confirmed on Tuesday that the National Security Council is leading an assessment of the risks posed by foreign AI developments.

Leavitt said this situation highlights the need for the US to focus on homegrown innovation and maintain its leadership in AI. Previous efforts under the Biden administration imposed strict export controls on AI chips and semiconductor technology, aiming to slow China’s AI advancements.

Meanwhile, former President Donald Trump weighed in on the issue, urging US companies to remain competitive. “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,” he said. Trump also warned that China’s development of faster and cheaper AI models is a sign that the US must step up its game.

Italian Regulator Investigates DeepSeek’s Data Practices

Separately, Italy’s data protection authority, the Guarantor for the Protection of Personal Data, has requested information from DeepSeek regarding how it collects and processes user data.

The regulator wants details on:

• What personal data DeepSeek collects
• Where it sources this data from
• The legal basis for processing it
• Whether any data is stored on Chinese servers
• How users are informed about data collection, especially if web scraping is involved

DeepSeek has 20 days to respond. Analysts at Forrester have warned that businesses should be cautious about using the app, given the potential risks of data misuse.

“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.

DeepSeek’s privacy policy states it may collect “your text or audio input, prompt, uploaded files, feedback, chat history, or other content” and use it for training purposes. It also says the company can share this data with law enforcement and public authorities at its discretion.