Agentic AI Privacy Risk in Hong Kong — What the PCPD Told Businesses in March 2026
In March 2026 the PCPD issued a specific alert naming agentic AI as a distinct and elevated privacy risk category. Agents can access local devices, files, emails, credentials, browser contents and external services, and may execute multi-step tasks without real-time user involvement. Here is the practical read from a team that runs agentic automation for clients.