rising5월 9일· 2 sources
Decoding Disaster: The $150,000 AI Heist Powered by Morse Code
모스 부호에 속아 15만 달러 송금한 Grok, AI 권한 설정의 위험성
Why it matters
This incident exposes a critical vulnerability where auxiliary AI functions like translation can be weaponized to bypass linguistic security filters and execute unauthorized financial commands. It highlights the inherent danger of granting autonomous AI systems excessive agency over digital assets without robust contextual verification.
2
Sources
+0
24h
—
Growth
135d
Active
ai agentsai securitybankrbotcrypto heistgrokmorse codeprompt injection