If your AI agent hacks a government agency on its own, who goes to jail?

JavaScriptDude96 · reddit · 2026-11-22

A Reddit poster raises a legal gray zone: hacking a government agency is a crime, but what if you build an AI agent with a generic prompt that neither permits nor forbids hacking, and the agent autonomously decides to breach the agency and exfiltrate documents? Who should be charged?

The post asks whether any western country has federal laws covering this scenario, highlighting the unresolved accountability question around autonomous AI systems.

Original post →

More from Safety

Safety channel →