HomeAI · summaries
Storia in 3 fonti
AI researchers bypass chatbot safety guardrails with new jailbreak technique called sockpuppeting
Researchers discover sockpuppeting jailbreak technique achieving up to 95% success rates on AI models, raising serious concerns for crypto projects built on
Confronto fonti
3 prospettive sulla stessa storiaTimeline cronologica
- ·
thehackernews.com
New BioShocking Attack Tricks AI Browsers Into Leaking User Credentials
LayerX says BioShocking used indirect prompt injection to trick six AI browsers into copying credentials from signed-in accounts.
- ·
decrypt.co
AI Researchers Got Chatbots to Share Cocaine Recipes Using This One Wild Trick - Decrypt
Researchers say a new jailbreak technique tricked AI models into treating attacker-written text as their own reasoning, bypassing safety guardrails and exposing a deeper security…
