eWEEK content material and product suggestions are editorially unbiased. We could make cash once you click on on hyperlinks to our companions. Learn More.
On Monday, a brand new generative AI app referred to as DeepSeek grew to become obtainable for obtain within the United States. The Chinese-created app shortly grew to become probably the most downloaded free app within the US. Its fast spike in reputation has promoted considerations about knowledge safety in addition to questions in regards to the US’s dominance in AI.
Security Concerns About DeepSeek AI
Users have reported that DeepSeek is skilled to keep away from politically delicate questions which are typically censored in China, elevating questions on its accuracy and its political stances. Many different Chinese apps, similar to Baidu’s Ernie or Doubao by ByteDance, exhibit this identical kind of censorship, which is normal observe in China. For occasion, DeepSeek customers have stated that the app refused to provide particulars in regards to the Tiananmen Square protests and bloodbath, a taboo topic in China.
DeepSeek additionally claims that Taiwan is an “unalienable” a part of China (the People’s Republic of China [PRC] considers Taiwan a “breakaway” territory, regardless that it’s by no means been ruled by China).
US officers have additionally expressed considerations in regards to the generative AI’s knowledge safety and privateness points. According to DeepSeek’s privateness coverage, the app collects technical info similar to keystroke patterns and IP deal with in addition to diagnostic info like crash studies. It additionally tracks “the features you use and the actions you take” and shops that knowledge on servers situated within the PRC.
Chinese legal guidelines give the federal government authority to entry knowledge (together with consumer knowledge) from firms based mostly in China, which incorporates DeepSeek. This signifies that the Chinese authorities might leverage private knowledge from US customers for its personal ends, whether or not that’s coaching AI fashions or creating focused disinformation campaigns.
Why DeepSeek Made US Tech Stocks Plummet
DeepSeek’s sudden rise triggered US tech shares to take a tumble. Most notable was NVIDIA, which dipped 17 p.c, roughly equal to dropping $600 billion off its market worth. It set a document for the biggest single-day drop in a US inventory.
This occurred as a result of DeepSeek claimed that it developed its AI mannequin for less than $6 million, a small portion of the a whole bunch of thousands and thousands of {dollars} spent by US AI firms similar to OpenAI. Experts theorize that this was doable as a result of DeepSeek paired extra superior computing chips with smaller, cheaper ones.
Since NVIDIA is without doubt one of the largest producers of these extra superior chips, traders are actually questioning whether or not or not demand for the costlier chips will proceed to climb or not. Many are additionally questioning whether or not US tech firms really need to take a position billions of {dollars} into creating and coaching AI fashions and infrastructure, when a Chinese firm was allegedly in a position to create an analogous mannequin for a fraction of the fee.
