Citing concerns over data leaks, the Chinese AI tool DeepSeek is facing backlash from the international community just days after its launch on January 20. The app has been removed from app stores in several countries and is under scrutiny from regulatory authorities.
DeepSeek faces global backlash, removed from app stores
The AI tool DeepSeek is facing international accusations of leaking user personal data (Illustration: Le Monde).
International Reactions
Immediately following its launch, DeepSeek attracted millions of downloads globally. However, this popularity came with concerns about data security. Italy’s Data Protection Authority (GPDP) issued warnings regarding the risk to users’ data, stating: “Millions of Italian users may have had their personal data collected by DeepSeek, and we are unclear about the intended use or storage locations for these information.”
Italian authorities demanded that DeepSeek provide detailed reports within 20 days. The app has since been removed from both Google Play and Apple’s App Store in Italy. Similar concerns were raised by regulators in Ireland, with Senator Malcolm Byrne emphasizing: “If the data is stored in Ireland or other EU countries, we can protect it. But if it’s stored in China, we cannot control it.”
Concerns Over Data Storage
DeepSeek launched two language models, DeepSeek-V3 and DeepSeek-R1, which are comparable in power to OpenAI’s ChatGPT. However, users must provide substantial personal information to access the service. According to DeepSeek’s privacy policy, data may be stored on servers located in multiple countries, including China, “as long as necessary.”
Cybersecurity firm Wiz (US) discovered a major data leak involving DeepSeek, revealing over a million lines of data, including chat logs and personal user information, exposed without proper security measures.
Lower Accuracy Than Competitors
In addition to security issues, DeepSeek also falls short in accuracy compared to its Western counterparts. NewsGuard reported that the AI chatbot achieved only 17% accuracy when answering current affairs questions, repeating false information in 30% of cases, and providing vague or unhelpful responses in 53% of cases.
Compared to chatbots like ChatGPT and Gemini, DeepSeek’s error rate is significantly higher, raising questions about the quality and reliability of the AI tool in providing accurate information to users.
Conclusion
DeepSeek is facing significant challenges to its reputation and trust from international users. Concerns over data security and information accuracy have led to its removal from app stores and stringent oversight from regulatory bodies. To regain user trust, DeepSeek needs to implement clear and transparent measures to protect user data and improve the quality of its information.
