Companies struggle to define normal behaviour for AI agents, report finds

A new report indicates that organisations are struggling to establish behavioural baselines for autonomous AI agents. Without a defined baseline of normal activity, distinguishing malicious or erroneous agent behaviour from legitimate operation is harder. The article does not name specific products, versions or recommended controls.

For security analysts · Written by AI from 1 source · Wrong?