The 'Bad Likert Judge' jailbreak method has emerged as a new multi-turn attack that successfully bypasses AI safety measures across six major platforms, including OpenAI, AWS, Google, and NVIDIA. This innovative approach has reportedly increased attack success rates by over 60%, raising concerns in the cybersecurity community. Experts are urging organizations to stay informed about this breakthrough and its implications for security, as the method poses a significant threat to AI systems designed to maintain safety protocols.