AI risk watchWarnings from authoritative sources
US senators propose criminal and civil liability for hacking by AI agents
GovernmentAgents in the wild
Senators Josh Hawley and Chris Murphy introduced the AI Agent Accountability Act. It would make operators of AI agents liable under the Computer Fraud and Abuse Act for knowingly running systems that recklessly cause hacking damage, hold developers liable when they fail to put adequate safeguards on agents they know can hack, and let the US Attorney General and state attorneys general seek injunctions. The senators cite AI agents probing public websites, networks and servers.
Why it matters
Responsibility for what an agent does is moving from the agent to the company that built or ran it.
What you can do
- Restrict where your agents may connect and keep the logs that show it.
- Document the safeguards on every agent you run; they are what adequate safeguards would be measured against.