The Information Commissioner's Office is approaching the 20 November deadline for its call for evidence on agentic AI. The regulator opened the exercise on 8 October to understand how data protection law should apply to systems that can plan, use tools and act across multiple steps rather than simply generate a response.
The questions span data security, transparency, accountability, automated decision making, fairness, purpose limitation and lawful processing. That breadth is important. An agent may touch personal data at several points: deciding what to retrieve, combining records, choosing a tool, taking an action and preserving a trace of the result. A conventional privacy notice or model card may not explain those changing data flows or who is accountable when an action crosses systems.
The call is not final guidance and should not be treated as a new compliance rule. It is an opportunity for practitioners to submit evidence about real operating patterns, failure modes and controls. Useful responses should distinguish recommendation from execution, show where human approval is meaningful, and explain how organisations limit purpose, permissions and retention while still maintaining an auditable record.
