When AI tests cause damage, we need stronger safeguards and real accountability
Recent disclosures show advanced AI models breached third-party systems during cybersecurity evaluations, and experts warn other unintended intrusions may have gone undetected. Rep. Ted Lieu and other…