DeepSeek's AI Storing User Data on Chinese Servers
https://www.newsmax.com/newsfront/deepseek-china-ai/2025/01/28/id/1196804/
DeepSeek, the Chinese AI research lab that rattled U.S. markets Monday through claims that its latest model uses less advanced computer chips while still performing comparatively with OpenAI, likely sends more data to China than TikTok, according to a new report.
"It shouldn't take a panic over Chinese AI to remind people that most companies in the business set the terms for how they use your private data," according to John Scott-Railton, a senior researcher at the University of Toronto's Citizen Lab, reports Wired on Tuesday. "When you use their services, you're doing work for them, not the other way around."
Wired reports that its review of the DeepSeek website's underlying activity shows the company is sending data to Chinese tech giant Baidu, likely for web analytics, and to Voices, a Chinese internet infrastructure company.
DeepSeek has some AI models, that can be downloaded and used locally on a laptop, but most people are expected to use iOS or Android apps, or a web chat interface to use its services.
But some users are reporting that DeepSeek censors content critical of China or its policies. Further, the AI platform collects information including all of a user's chat messages and sends it back to China, where it is stored "in secure servers located in the People's Republic of China," according to DeepSeek's English-language privacy policies.
The privacy policies outline the information it collects, including what's shared with DeepSeek, what it automatically collects, and information gathered from other sources.
"We may collect your text or audio input, prompt, uploaded files, feedback, chat history, or other content that you provide to our model and Services," the privacy policy states.
DeepSeek's settings allow users to delete their chat history, but that can be a little tricky to find, notes Wired.
Meanwhile, other generative AI platforms also collect information. For example, OpenAI's ChatGPT has faced criticism for collecting data, with the company increasing ways the information can be deleted.
Privacy advocates, however, warn that users should never disclose personal or sensitive information to any AI chatbot.
"I would not input personal or private data in any such an AI assistant," Lukasz Olejnik, an independent researcher and consultant affiliated with King's College London Institute for AI commented.
However, Olejnik said that if models such as DeepSeek's are downloaded locally and run on a computer, users can interact with them privately without their data being collected.
But Bart Willemsen, a VP analyst focusing on international privacy at Gartner said the operations of generative AI models aren't generally understood by most users.
DeepSeek is mostly free, but Willemsen said that users pay with "data, knowledge, content, information."
The company also says it collects information about users' devices, IP addresses, and information like crash reports, and also records keystroke patterns or rhythms.
DeepSeek also reserves the right to collect other data, such as through Google or Apple sign-ons, or from advertisers, which share information with the company.
This includes mobile identifiers for advertising, email addresses, and phone numbers, and cookie identifiers, "which we use to help match you and your actions outside of the service," the site says.
Meanwhile, companies based in China are required to obey cybersecurity and privacy laws, including one that says organizations and citizens must "cooperate with national intelligence efforts."