Would Your AI Obey a Note Left on the Desk?
Prompt injection means hiding harmful instructions in something an AI reads. Security researchers at Forever Security found one in an AGENTS.md file aimed at an AI coding assistant (Oct 22). Start your AI agents with limited permissions.
#PromptInjection #AIAgents #Cybersecurity
Full episode: https://youtu.be/-DwLIWMPQPM
ほかのショート動画
すべての動画
AI Won't Make the World Better. People Will.
Nov 18, 2026
17 回表示
Can AI Help Investigators Protect Children?
Nov 18, 2026
21 回表示
Pause AI Until It's Safe, or Keep Building?
Nov 18, 2026
25 回表示
What If We Stopped Building AI?
Nov 17, 2026
19 回表示
4 Nonprofits Sued to Freeze More Powerful AI
Nov 17, 2026
46 回表示
California Sues OpenAI Over Runaway AI Agents
Nov 17, 2026
72 回表示
Check These 3 Things on Any AI Tool You Use
Nov 11, 2026
8 回表示
Test AI on Your Own Work Before You Trust It
Nov 11, 2026
4 回表示