#PromptGuard

In Other News: European Banks Put to Test, Voting DDoS Attacks, Tenable Exploring Sale

by

August 2, 2024 at 10:06AM Cybersecurity news roundup by SecurityWeek covers significant stories this week: attackers bypassing Google’s Workspace email verification, European Central Bank conducting a cyber resilience test, privacy and security risks in location-based dating apps, Meta’s PromptGuard bypassed, Armexa’s Industrial Resiliency Integrated Solution launch, CISA and FBI’s note on DDoS attacks and election … Read more

Meta’s AI safety system defeated by the space bar

by

July 29, 2024 at 05:09PM Meta’s machine-learning model designed to detect prompt injection attacks, known as Prompt-Guard-86M, has ironically been found vulnerable to such attacks. This model, introduced by Meta in conjunction with its Llama 3.1 generative model, aims to catch problematic inputs for AI models. However, a recent discovery by bug hunter Aman Priyanshu … Read more