Comment by Andrew Burt

AI agents can behave a lot like malware, acting autonomously and causing harm if left unchecked. Organizations need reliable kill switches that can shut agents down when they misbehave. Unverified source (2026)