ICO sets out plan for safe AI innovation in the UK

TL;DR:

• The Information Commissioner’s Office has published a plan setting out how it will enable safe AI innovation across the UK public and private sectors.
• Priorities for 2026/27 include a dedicated AI code of practice, specific guidance on agentic AI, and new support to help consumers navigate personalised systems.
• It is the latest UK regulator to choose guidance over fresh rules — a now-familiar pattern across the post’s data, finance and competition watchdogs.

Britain’s data protection regulator has set out how it intends to balance AI innovation with public trust, responding to a January request from the Technology and Business Secretaries to detail its safety strategy. The resulting plan leans on clarity rather than new restrictions, promising developers more certainty on how data protection law applies to AI as the technology moves deeper into everyday products.

Guidance over new rules

The ICO says the next 12 months will focus on building consumer trust and giving businesses “greater regulatory certainty” on AI development and deployment. Concrete commitments include a dedicated AI code of practice, guidance specifically on agentic AI — systems that act with growing autonomy — and support to help consumers handle increasingly personalised services. A fuller 2026/27 workplan is due in an upcoming strategy document.

The approach fits a wider UK regulatory posture. It echoes the FCA’s decision to lean on existing frameworks rather than write new AI rules and the government’s AI Growth Labs, with legal services first in line. The consistent message from Whitehall is that Britain will compete on regulatory clarity and speed, not on a heavy new rulebook — a bet that certainty itself is a growth lever.

Looking forward

The plan’s value will hinge on delivery. Promised codes and agentic AI guidance address two of the thorniest questions facing UK organisations, but only once published in usable form. For data and compliance teams, the practical signal is to expect firmer expectations on AI accountability over the coming year — and to prepare for an agentic-AI code rather than wait for it.