JU has not yet reached whether a Chinese AI company is a game changer or a part of a short NVIDIA and other technical stocks by hedge fund’s parent company. 。 In either case (probably both?), Deepseek and its large language models made several large waves. And now, it’s catching the data protection watchdog.
Today, the Irish Data Protection Committee has confirmed with TechCrunch that the Irish citizen’s data was processed by the company to DeepSeek’s note to DeepSeek. “The Data Protection Committee (DPC) is described in DeepSeek, which requires information about data processing implemented in connection with Irish data,” said a spokeswoman. They refused to answer further questions.
Letters from Irish DPAs were sent within 24 hours after Italian data protection watchdogs sent similar notes to the company. Deepseek has not yet responded to the public demands. However, the mobile app is no longer displayed in both Italian Google and Apple app store.
The movement in Italy seemed to be the first big movement from such watchdogs, as DeepSeek has recently become a viral. EUROCONSUMERS, a coalition of European consumer groups, has complained to Italian data protection agencies related to how DeepSeek is processing personal data on GDPR, a framework for European data protection regulations.
Italian DPA today confirmed today what I wrote in Deepseek at the information request. “Lissio di Millioni di Persen in Italy,” he says. (“Millions of Italian data are in danger.”) DeepSeek has a 20 -day response.
Two important details of Deepseek that many people have noticed are that services are created and operated by China. According to the privacy policy, this includes the information and data that DeepSeek collects and stores stores. This is also in your home country.
DeepSeek is also briefly stated that when Deepseek is used to transfer data from the country where Deepseek is used to China, he moves data in accordance with the requirements of the Applied Data Protection Law.
However, EUROCONSUMERS -Last year, it is an organization that filed a successful lawsuit against GROK about how to train AI to train AI, and Italian DPA wants more detailed details.
Hangzhou Deepseek If you work on artificial intelligence and Beijing Deepseek Artificial Intelligence, Italian DPA, which includes information that is collected, source, and any purpose, including information used to train AI systems. He said he wanted to know legal things. The foundation is for processing. It is also necessary for the details of the server in China.
Furthermore, in the information request, “when personal data is collected through a web scraping activity”, “Registered person and users who are not registered in the service are processed, or about their treatment. I want to know if they are notified. “
NEWS Outlet Mlex emphasizes that EUROCONSUMERS has no details about how to protect or restrict minors to that service, from age verification to how to handle minor data. I have stated.
(DeepSeek’s age policy is not for users under the age of 18, but we do not provide it. DeepSeek is between 14 and 18 years old. Proposes to read a privacy policy.
EUROCONSUMERS and Italian watchdogs represent the first effort to move with DeepSeek. They may not be the last, but the follow -up may not be so quick.
Today, DeepSeek was a major topic at a press conference at the European Commission. THOMAS REGNIER, a spokesman of the Tech Souvereignty committee, was asked if he was more concerned at the European level than the security, privacy, and censorship -related Deepseek. However, at the moment, the main messages seemed to be as follows. It’s too early to say something about the survey.
“The services provided in Europe respect our rules,” Regnier pointed out in questions about data privacy, and the AI method is applied to all AI services provided in this area. I added.
He refused to say whether Deepseek respected these rules in EU estimation. Later, he asked if the censorship of an app on a politically sensitive topic in China violated the free rules of European speech and whether it was worth a survey. “These are very early stages, and I’m not talking about the investigation yet,” Regnier immediately said. “Our framework is enough to work on potential problems if you are here on potential problems.”
TechnoCrunch received a similar answer to the question sent to the ICO to the ICO in the UK: Deepseek is, in effect, the same scrutiny as other GENAI developers. But there is no further action yet.
“The generated AI developers and developers need to make sure that people have information that can be accessible, simple and easy to access in terms of use of personal data, so that people can exercise information rights. I need to have a clear and effective process to do it, “said a spokeswoman. “We will continue to be involved with stakeholders to promote effective transparency measures without avoiding the action of regulatory authorities.
On the other hand, are there any new measures for regulatory authorities around the area, such as copyright and IP protection?
Many people are surprised at how DeepSeek’s existence itself is challenging the actual cost of training and operation of LLM or generated AI services. Properties for the use of chips, data centers, and energy consumption.
But recently, some have begun to question all of them. Microsoft and Openai say that there is evidence that “distillation” from its own models has been partially trained. If it is proven that it is true, it will be eerie -irony -a number of legal dramas that are said to have some LLM builders see intellectual property and copyright. Considering that other dramas are swirling.
Contact Deepseek about Italian DPA complaints, and update this post when more information is available. Until then, the Deepseek app has been drawn out of the main Italian app store, but it seems to be online and live in Japan.
Details of Italian regulations, legal issues, and the status of services have been updated.