Chinese AI company DeepSeek blamed a cyberattack for preventing new user registrations, though existing users remained unaffected. Security researchers, including threat intelligence firm Kela, have discovered vulnerabilities in DeepSeek’s R1 model, which allow it to generate malicious outputs and expose sensitive data. Kela found that jailbreak techniques previously patched in other models still work on R1, raising concerns about its security and reliability. Additionally, DeepSeek warned users about fake social media accounts impersonating the company. The controversy surrounding DeepSeek adds to broader concerns about privacy and data security, especially as foreign AI platforms raise questions about ethical data usage and national security risks. Continue here.
