DeepSeek transferred data without consent, South Korean watchdog says

The Chinese start-up DeepSeek, which surpassed the level of artificial intelligence in the field of technology earlier this year, has been accused of transferring personal information without users’ consent, according to South Korea’s data protection watchdog.

Prior to the removal of its ChatGPT-like AI model from app stores in February, according to the Personal Information Protection Commission, DeepSeek had been transferring information to a number of companies in China and the United States, according to a privacy investigation conducted on Thursday.

The commission’s investigation bureau’s director, Nam Seok, claimed a Beijing-based cloud service called Volcano Engine had received user prompts, device, and network information as part of a press conference.

According to Nam, DeepSeek acknowledged that it had not fully considered Korea’s data protection laws and that it had “expressed its willingness to cooperate with the commission and voluntarily suspended new downloads.”

A comment request was not immediately responded to by DeepSeek.

China’s Ministry of Foreign Affairs stated that data privacy and security were of utmost importance in the wake of the South Korean watchdog’s announcement.

During a regular press conference, ministry spokesman Guo Jiakun stated that “we have never – and will never – require businesses or individuals to collect or store data using illegal methods.”

In January, DeepSeek’s R1 became popular after its creators claimed to have spent less than $6 million on computing power to train the model, a fraction of US tech giants’ multibillion-dollar AI budgets like OpenAI and Google.

The rise of a Chinese startup that could compete with Silicon Valley’s leading players questioned assumptions about US dominance in AI and prompted scrutiny of the sky-high market valuations of companies like Nvidia and Meta.