DeepSeek AI Platform Found to Have Zero Security Controls, Exposing Users to Major Risks
• 1 min read
A joint investigation reveals DeepSeek's AI model is completely vulnerable to jailbreak attacks, while the company left user data exposed in an unsecured database. Security experts warn organizations to immediately block access to DeepSeek services due to unacceptable security risks.
DeepSeek AI Model Shows Critical Security Flaws in Cisco Testing
• 1 min read
Cisco's comprehensive testing reveals DeepSeek R1 AI model failed to block any harmful prompts, achieving a concerning 100% attack success rate. The findings raise serious security questions as the model becomes available on major cloud platforms.