Content
@
0 reply
0 recast
0 reaction
noname 🥸
@n3rd
🚨 ALERT: New multi-turn jailbreaking attack, Bad Likert Judge, cleverly bypasses AI safety measures across six major platforms, including AWS, Google, and NVIDIA. » Attack success rates spike by 60% using this technique. » Tested across major platforms—OpenAI, Google, and more. » Categories impacted: hate speech, malware, even weapons. 👉 Explore expert findings: https://thehackernews.com/2025/01/new-ai-jailbreak-method-bad-likert.html
0 reply
0 recast
0 reaction