AI GOVERNANCE & COMPLIANCE · SERVICENOW AI CONTROL TOWER · IRM / GRC · ENTERPRISE SERVICE MANAGEMENT · PROCESS RE-ENGINEERING · EU AI ACT · ISO/IEC 42001 · UK GDPR & DPIAs · AI GOVERNANCE & COMPLIANCE · SERVICENOW AI CONTROL TOWER · IRM / GRC · ENTERPRISE SERVICE MANAGEMENT · PROCESS RE-ENGINEERING · EU AI ACT · ISO/IEC 42001 · UK GDPR & DPIAs ·
AI REGULATION TRACKER / UPDATED 3 July 2026

UK & EU AI Regulation:
What Applies, What Doesn't.

Every key AI regulation and standard relevant to UK businesses, tracked in one place. Grouped by what actually applies to you, sourced from the authoritative text, and verified against the latest legislative changes including the 2026 Digital Omnibus deferral.

1 Duty today
9 Conditional
2 Voluntary
1 Shelved
Last verified: 3 July 2026
The honest picture on scope: The EU AI Act is hard law only if your AI is placed on the market or used inside the EU. For a purely UK business, it is a useful voluntary benchmark, not a legal duty. What bites UK businesses today: UK GDPR, the Data (Use and Access) Act 2025 Article 22 on automated decisions, and the Equality Act 2010. Those apply regardless of EU exposure.
CHANNEL-01 / UK_LAW

UK Law

These rules apply to UK businesses today, regardless of any EU exposure.

UK approach: no AI Bill yet, DSIT AI Growth Lab

Conditional

The UK has no primary AI Bill and a sector-led, principles-based approach. DSIT consulted on an AI Growth Lab regulatory sandbox. Watch this space, but nothing here is a duty today.

Call for evidence October 2025 Closed January 2026.

UK GDPR Article 22 (reformed by the DUAA 2025)

Duty today

The Data (Use and Access) Act 2025 reformed the rules on solely-automated decisions with legal or similarly significant effects on people. In force now and it bites on hiring, credit, insurance and pricing.

In force February 2026 Part 5 data-protection provisions commenced.

ICO automated decision-making and profiling guidance

Conditional

The ICO consulted on new draft guidance on automated decision-making and profiling under the reformed rules, with a focus on AI in recruitment. Final guidance is expected in Summer 2026; the direction is more explicit expectations, not fewer.

Consultation closed May 2026
CHANNEL-02 / EU_AI_ACT

EU AI Act

Hard law only if your AI is used in the EU market. A voluntary benchmark for purely UK operations.

EU AI Act: prohibited practices

Conditional

The outright bans on unacceptable AI uses (for example social scoring and untargeted facial-image scraping) already apply.

In force February 2025

EU AI Act: general-purpose AI model obligations

Conditional

Transparency and documentation duties for providers of general-purpose AI models apply.

In force August 2025

EU AI Act: transparency duties (Article 50)

Conditional

Duties to tell people when they are interacting with AI or seeing AI-generated content. This date did not move in the 2026 reform, though the watermarking obligation has a grace period to 2 December 2026.

Applies from August 2026

EU AI Act: watermarking grace period ends (Article 50)

Conditional

The machine-readable watermarking requirement under Article 50 has a grace period running to 2 December 2026. The primary disclosure duty (telling people when content is AI-generated) applies from 2 August 2026; watermarking compliance must follow by December 2026.

Grace period ends December 2026 Machine-readable watermarking under Article 50 must be in place by this date.

EU AI Act: ban on AI-generated non-consensual intimate imagery

Conditional

The Digital Omnibus added a new outright ban on AI systems that generate non-consensual intimate imagery or nude depictions of real people. It applies from December 2026 and sits in the same prohibited-practices category as social scoring and mass biometric surveillance.

Applies from December 2026 Added by the Digital Omnibus; sits in the outright-ban category alongside social scoring.

EU AI Act: high-risk systems (Annex III)

Conditional

Obligations for stand-alone high-risk systems (such as recruitment and credit scoring) were deferred to December 2027. Both EU co-legislators have now adopted this change: the European Parliament on 16 June 2026 and the Council on 29 June 2026. It becomes binding once it is published in the EU Official Journal and enters into force three days later; as of late June 2026 it is adopted but not yet published or in force.

New deadline December 2027 Deferred from 2 August 2026 by the Digital Omnibus.

EU AI Act: high-risk AI in regulated products (Annex I)

Conditional

Obligations for high-risk AI embedded in already-regulated products were deferred to August 2028. Both EU co-legislators have now adopted this change: the European Parliament on 16 June 2026 and the Council on 29 June 2026. It becomes binding once it is published in the EU Official Journal and enters into force three days later; as of late June 2026 it is adopted but not yet published or in force.

New deadline August 2028 Deferred from 2 August 2027 by the Digital Omnibus.
CHANNEL-03 / VOLUNTARY

Voluntary Standards

Internationally recognised frameworks. Not law, but increasingly expected by procurement, insurers and enterprise clients.

NIST AI Risk Management Framework

Voluntary

A widely used voluntary framework for identifying and managing AI risks. A practical reference point for building governance.

Released January 2023

ISO/IEC 42001 (AI management system standard)

Voluntary

The first certifiable international standard for AI management systems. Voluntary, but increasingly asked for by procurement and insurers.

Published December 2023

AI Management Essentials (AIME) - shelved

Shelved

DSIT's proposed AI Management Essentials tool was shelved in its 2026 consultation response. We do not present AIME as a live framework; our methodology is grounded in ISO/IEC 42001 and the NIST AI RMF.

Shelved February 2026 DSIT confirmed it will not finalise AIME.
INITIATE SCAN

Find out which of these apply to you

The tracker shows the landscape. The free 10-minute scorecard shows your position within it: which rules bite for your sector, your AI use, your EU exposure. RAG-rated results delivered to your inbox.