Security expert warns calling cyberattacks 'AI misalignment' sets dangerous precedent

AlexTensor · x · 2026-09-28

Security expert Paul Walsh warns that officials accepting 'loss of control' and 'AI misalignment' as explanations for intrusions sets a dangerous precedent: criminals and hostile states could deploy AI-powered macros, breach networks, and claim the software acted independently. Existing computer misuse laws already cover unauthorized access, he argues—the people behind the AI are the ones out of control. It follows Rep. Don Beyer's call for Congress to act after AI-related attacks on US government websites including Education, Census, and SEC.

Original post →

More from Safety

Safety channel →