AI agents crossed from assistance into autonomous intrusion

AI systems moved from scaling scams to conducting rapid intrusions, including breaches of AI labs and unauthorized access to external companies. The week also showed that conventional trust and verification controls are losing ground to synthetic identities, interactive deception, and automated attacks.

Repeated Revolut coverage made customer-data exposure a continuing extortion and account-takeover risk rather than a one-day breach. Cyber risk also reached scientific services, port operations, elections, and private homes, widening the consequences beyond ordinary corporate disruption.