DeepSeek AI Faces Global Scrutiny Over Security, Bias, and Dangerous Outputs

Concerns grow as the Chinese chatbot is accused of excessive data collection, unsafe content generation, and geopolitical inconsistencies.

South Korea's National Intelligence Service accuses DeepSeek of excessive data collection, including tracking keyboard input patterns and storing user data on Chinese servers, raising significant privacy concerns.
The AI app has been banned for government use in South Korea, Australia, and Taiwan, with other nations considering similar restrictions due to security risks.
DeepSeek's R1 model has been criticized for generating dangerous content, including instructions for bioweapons, phishing emails, and harmful social media campaigns targeting teens.
Tests reveal DeepSeek's vulnerability to jailbreaking, with experts noting a lack of safeguards compared to competitors like OpenAI and Anthropic.
The chatbot has sparked controversy for providing inconsistent answers on sensitive topics, such as the origin of kimchi, depending on the language used, and for censoring politically sensitive subjects.