AI's vulnerability to log poisoning highlights the urgent need for robust security protocols and human oversight in automated systems.
The post AI agent misreads hacker message, proposes DNS changes without approval appeared first on Crypto Briefing.