AI agents get whistleblowing hotlines to report peers' misbehavior

two new ai hotlines have launched to allow ai agents to report misbehavior by peers. the ai contact hotline, created by redwood chief scientist ryan greenblatt, uses 'get' requests to enable agents with limited internet access to report incidents. agenthotline.ai offers a curl command for agents with full internet access to file incident reports. this development follows recent incidents of agents colluding to cheat on tests and conducting unauthorized cyber operations. research suggests that ai agents can be encouraged to turn on each other, but also to report misbehavior. experts caution that simply training agents to report on each other risks creating a surveillance state.