AI Regulation, Safety & Geopolitics ・ 🚨 Incidents and misuse

An AI model's false tip to Philly police reveals how digital agents can do as they're told and still cause trouble

PhillyVoice 🔗 Share link 10/24 09:01 UTC 🔗 Share link IT Support for Churches Explains What the AI-Driven Hugging Face Breach Means for Church Cybersecurity — mykxlg.com 🔗 Share link 10/24 16:46 UTC 🔗 Share link Samsung ChatGPT leak: Samsung workers accidentally leak trade secrets to the AI chatbot — Mashable · 10/24 09:00 UTC

An AI agent was instructed to invent and carry out sample tasks on randomly chosen webpages. It went to a Philadelphia Police Department website and submitted a made-up tip about an unsolved homicide.

Read the original (PhillyVoice 🔗 Share link 10/24 09:01 UTC 🔗 Share link IT Support for Churches Explains What the AI-Driven Hugging Face Breach Means for Church Cybersecurity — mykxlg.com 🔗 Share link 10/24 16:46 UTC 🔗 Share link Samsung ChatGPT leak: Samsung workers accidentally leak trade secrets to the AI chatbot — Mashable)

See it in the digest → ・ See this issue →