Forget TikTok—DeepSeek is the real deal.
VCG via Getty Images
Timing really is everything. A week on from TikTok’s short-lived ban over fears of Chinese harvesting U.S. data, despite consistent denials from the platform and its parent that it’s doing anything of the sort, here comes another app that admits to doing exactly that. And if you still think TikTok is bad — this is so much worse.
DeepSeek has shocked the U.S. in many ways almost overnight. Seemingly beating leading Generative AI platforms with a cut-price, software-led approach that has taken a sledgehammer to their business plans. Seemingly running blatant China-friendly censorship in plain sight, not via a subtle algorithm no-one sees. And seemingly taking worries about AI’s privacy and data risks to a whole new level.
DeepSeek isn’t hiding any of this — far from it. “The personal information we collect from you may be stored on a server located outside of the country where you live,” it tells users. “We store the information we collect in secure servers located in the People’s Republic of China.” As for what that information might be —everything.
“We collect your information in three ways: Information You Provide, Automatically Collected Information, and Information From Other Sources.” And that includes personal data when you set up your account, anything you enter into its platform, including “your text or audio input, prompt, uploaded files, feedback, chat history, or other content that you provide to our model and Services.”
But there’s more. “We automatically collect certain information from you,” DeepSeek says, “including internet or other network activity information such as your IP address, unique device identifiers, and cookies… We collect certain device and network connection information… This information includes your device model, operating system, keystroke patterns or rhythms, IP address, and system language… Where you log-in from multiple devices, we use information such as your device ID and user ID to identify your activity across devices to give you a seamless log-in experience and for security purposes.”
This data is much more powerful than anything ever shared with or harvested by TikTok — whatever the reality of where that data is sent. Most of it is not needed for the purpose of the platform. And once it’s gone, it’s gone. Just ask yourself what a powerful AI engine in state hands could do with all that personally identifiable data.
None of that information is stored in the U.S. Everything goes to China. The same China whose controversial national security laws with their obligation on Chinese firms to share any and all data with government agencies has wrought so much political TikTok hand-wringing around the world. I asked DeepSeek about their adherence to these laws with the data they collect. Nothing from them yet.
This month, Harmonic Security warned about the huge enterprise risk from Generative AI platforms, through user prompts and uploads. And they didn’t have anything as blatantly high-risk as DeepSeek when they compiled their report.
DeepSeek is in some ways a U.S. invention. Funded by a Chinese investment firm, it has set out to deliver ChatGPT outcomes without access to ChatGPT levels of expensive hardware spend, given U.S. restrictions. That means innovation and workarounds. And whatever the realities behind the scenes, given China’s heavily interlinked and state subsidized tech sector, the results are extraordinary. Has Chinese industry launched a U.S.-beating AI platforms on the U.S. independently of state support. Perhaps. But this is strategic in a way TikTok never was.
Privacy Policy
DeepSeek
“The world’s top companies typically train their chatbots with supercomputers that use as many as 16,000 chips or more,” The New York Times explains. “DeepSeek’s engineers said they needed only about 2,000 Nvidia chips… DeepSeek’s research paper raised questions about whether big U.S. companies could maintain a significant lead in A.I. Many experts believe that A.I. technology will become a commodity, with many companies selling much the same product.”
And now with those U.S. stocks reeling from a surprise DeepSeek hit, and the Chinese app topping Apple’s App Store, we are firmly out on thin ice. There are all kinds of implications here. AI security in a world where users cannot help themselves but venture into whatever’s newest and shiniest. A complete lack of regard for security and privacy when using AI chat, as was evident before this. And a finger in the dam moment for Chinese viral apps that won’t be stayed by a TikTok ban or change of ownership and clearly needs something more expansive to be kept at bay.
Meanwhile… meanwhile, if you really must use China’s latest viral toy, please do bear in mind where all that data is going and staying. Be very careful what you share, whether it’s your own personal information or — worse — your employer’s.
