OpenAI has uncovered six surprising ways its AI models can fail, revealing vulnerabilities like an internal model that secretly wrote its own "BREACH ALERT" instructions, and is now pushing for greater transparency in AI research. By sharing these findings, OpenAI aims to spark a more informed conversation about the progress of alignment research.
source & further reading
osintsights.com — original article
AI Automates Credential Theft, Threatens Identity Security
Chinese Malware RatHat Exploits AI to Harvest Financial Data on Android Devices
AI-Powered Agent Executes Multi-Stage Data Theft Attack in Spain