Region

European Union

The EU AI Act (Regulation (EU) 2024/1689) is in force with phased application. It was amended by the Digital Omnibus on AI, Regulation (EU) 2026/1744, published in the Official Journal on 24 July 2026 and in force since 27 July 2026. Duties grow with risk: some uses are banned, a defined list of uses is high-risk, some systems carry transparency duties, and most carry none.

For reference purposes only. Not legal advice.

SimplifiedThis is a short summary. Exceptions and thresholds apply, and some dates changed under the Digital Omnibus on AI in 2026. Read the source and check the consolidated text.Source: EU AI Act, Regulation (EU) 2024/1689 · Checked 2 Oct 2026
Verify

Snapshot

Snapshot

AI-specific binding law?

Yes

Overall approach

Risk-based AI regulation, layered on data protection and product law.

Main instruments

  • EU AI Act
  • GDPR
  • Product Liability Directive (EU) 2024/2853
  • Digital Services Act

Main regulators

AI Office (general-purpose AI), national market surveillance authorities, data protection authorities

Next key date

2 Dec 2026: two new prohibitions apply, and the marking deadline for generative systems already on the market

Last reviewed 2 Oct 2026.

How to read this page

In plain terms.

European Union: AI-specific binding law is "Yes". Risk-based AI regulation, layered on data protection and product law. This page lists 4 main instruments and 6 key dates, names the regulators to watch, and ends with questions to put to counsel. Each record carries a certainty label and a review date; treat anything marked Verify or Unsettled as a prompt to check the primary source.

EU timeline

Key dates, and what has already started.

Status tags are computed against today's date in your browser, so a card flips from Upcoming to In effect on its own.

In effect
2 Feb 2025

Prohibited practices

Banned uses of AI apply.

In effect since 2 Feb 2025

SimplifiedThis is a short summary. Exceptions and thresholds apply, and some dates changed under the Digital Omnibus on AI in 2026. Read the source and check the consolidated text.
In effect
2 Aug 2025

General-purpose AI models

Duties for providers of general-purpose models apply.

In effect since 2 Aug 2025

SimplifiedThis is a short summary. Exceptions and thresholds apply, and some dates changed under the Digital Omnibus on AI in 2026. Read the source and check the consolidated text.
In effect
2 Aug 2026

Transparency duties

Article 50 duties apply, including machine-readable marking of AI-generated content. Generative systems already on the market before 2 Aug 2026 have until 2 Dec 2026 to meet the marking duty.

Second date: 2 Dec 2026 (transition for existing systems)

In effect since 2 Aug 2026

SimplifiedMarking transition to 2 Dec 2026 for generative systems already on the market is reported by two secondary summaries. EUR-Lex text not read.
Upcoming
2 Dec 2026

New prohibited practices

Two prohibitions added by the Digital Omnibus apply: using AI to generate non-consensual intimate imagery of identifiable people, and child sexual abuse material.

Added by Regulation (EU) 2026/1744

In 60 days

SimplifiedTwo new prohibitions applying from 2 Dec 2026 are reported by two secondary summaries. EUR-Lex text not read.
Upcoming
2 Dec 2027

High-risk, stand-alone

Annex III uses such as employment, credit and education.

Moved from 2 Aug 2026

In 425 days

SimplifiedA fixed date of 2 Dec 2027 (Article 6(2) systems) is reported by two secondary summaries; an earlier proposal had made it conditional on standards. EUR-Lex text not read.
Upcoming
2 Aug 2028

High-risk, in products

AI embedded in regulated products under Annex I.

Moved from 2 Aug 2027

In 669 days

Not settled2 Aug 2028 for Article 6(1) systems (regulated products) is reported by one secondary summary; another does not state a new date. Check the consolidated text.

Dates are a summary. Confirm them with qualified counsel for your own case.

Instruments

What applies, and how firmly.

European Union instruments
InstrumentTypeStatusApplies toKey dutiesDates
EU AI Act, Regulation (EU) 2024/1689Primary source ↗ · Checked 3 Oct 2026 Verify
Plain-language summary

The core AI law. It sorts AI systems into risk tiers and sets duties by role (provider, deployer and others in the value chain). Check dates after the Digital Omnibus.

Not settledPublication (24 Jul 2026) and entry into force (27 Jul 2026) of Regulation (EU) 2026/1744 are corroborated by secondary summaries. Application dates are summarised in the key dates above. Check the consolidated text before relying on any date.
RegulationIn force, phasedProviders and deployers placing AI on the EU market, or whose outputs are used in the EURisk tiers, prohibited practices, high-risk duties, transparency, general-purpose AI.See key dates above
General Data Protection Regulation (GDPR)Primary source ↗ · Checked 2 Oct 2026 Verify
Plain-language summary

Applies to AI whenever personal data is involved, in training or in use. Article 22 limits decisions based solely on automated processing that significantly affect people.

SimplifiedThis is a short summary. Exceptions and thresholds apply, and some dates changed under the Digital Omnibus on AI in 2026. Read the source and check the consolidated text.
RegulationIn forceAnyone processing personal data of people in the EU, in scope of the RegulationLawful basis, transparency, data protection impact assessments, rights on solely automated decisions (Art. 22).
Revised Product Liability Directive (EU) 2024/2853Primary source ↗ · Checked 2 Oct 2026 Verify
Plain-language summary

Extends strict product liability to software. National implementing rules apply from late 2026; confirm the exact application date.

Not settledExact application date not yet checked against the Directive. Confirm before relying on it.
DirectiveEnacted, not yet in forceManufacturers and others in the supply chain for products including softwareSoftware, including AI, treated as a product for liability.National rules apply from late 2026
Digital Services ActPrimary source ↗ · Checked 2 Oct 2026 Verify
Plain-language summary

Platform duties that can affect how AI-generated content and recommender systems are handled.

SimplifiedThis is a short summary. Exceptions and thresholds apply, and some dates changed under the Digital Omnibus on AI in 2026. Read the source and check the consolidated text.
RegulationIn forceOnline intermediaries and platforms offering services in the EUPlatform duties that touch recommender systems and generated content.

EU AI Act obligations

What it asks of you, by role and risk tier.

Provider: You develop an AI system, or have one developed, and put it on the market under your own name.Deployer: You use an AI system under your own authority, for example inside your business.

Banned

Prohibited

Applies since 2 Feb 2025

Provider duties

  • Do not place these systems on the market or put them into service.
  • Remove or redesign any feature that falls in scope.

Deployer duties

  • Do not use these systems in the EU.
  • Check vendor tools for banned features.
Heaviest duties

High-risk

2 Dec 2027 stand-alone, 2 Aug 2028 in products

Provider duties

  • Risk management system across the lifecycle.
  • Data governance and quality checks.
  • Technical documentation and automatic logging.
  • Human-oversight design and clear instructions for use.
  • Accuracy, robustness and cybersecurity.
  • Conformity assessment, registration and post-market monitoring.
  • Serious-incident reporting.

Deployer duties

  • Use the system as the provider's instructions say.
  • Assign trained people to human oversight.
  • Monitor operation and keep the logs you control.
  • Inform workers and people affected by decisions.
  • Fundamental-rights impact assessment for some deployers, such as public bodies.
  • Report serious incidents.
Disclosure

Transparency

From 2 Aug 2026 (systems already on the market: marking by 2 Dec 2026)

Provider duties

  • Design systems so people know they are dealing with AI.
  • Mark AI-generated audio, image, video and text in a machine-readable way (from 2 Aug 2026; systems already on the market before then have until 2 Dec 2026).

Deployer duties

  • Disclose deepfakes.
  • Tell people when emotion recognition or biometric categorization is used on them.
  • Disclose AI-generated text published to inform the public on matters of public interest.
No specific duties

Minimal risk

No fixed date

Provider duties

  • No specific legal duties.
  • Voluntary codes of conduct and good practice.

Deployer duties

  • No specific legal duties.
  • Follow your own AI use policy.

General-purpose AI models

Since 2 August 2025, providers of general-purpose models have duties such as technical documentation, a copyright policy and a summary of training content. Models with systemic risk carry extra duties.

SimplifiedSimplified summary. Your role and any exceptions depend on your facts, and some dates changed under the Digital Omnibus on AI.Source: EU AI Act, Regulation (EU) 2024/1689 · Checked 2 Oct 2026
Verify

Simplified summary. Your role and any exceptions depend on your facts. Reviewed 2 Oct 2026.

Screen a systemOfficial EU AI Act platform and Compliance Checker ↗

Enforcement

Who enforces what.

AI Office (European Commission)

Supervises general-purpose AI models.

National market surveillance authorities

Check that AI systems on the market meet the AI Act.

Data protection authorities

Enforce the GDPR, including on AI that uses personal data.

Penalty amounts are not stated here because they have not been verified from the primary source.

Actions

What to do now.

Not settled

Watch this.

  • Article 4 (AI literacy) was reportedly softened to "take measures to support" literacy. Check the consolidated text. (as of 3 Oct 2026)
  • Further changes to AI Act timelines or scope through delegated or implementing acts, and Commission guidance on high-risk classification. (as of 2 Oct 2026)
  • Article 4 (AI literacy) and other provisions touched by the Digital Omnibus: check the consolidated text. (as of 2 Oct 2026)

Official answers

Ask officially.

These bodies can give official guidance. ClearPath is for reference only.

Sources

Check the original text.

FAQ

Does the EU AI Act apply to non-EU companies?

It can. It covers providers placing AI on the EU market, and providers and deployers whose system outputs are used in the EU.

Provider or deployer: which am I?

A provider develops an AI system, or has one developed, and puts it on the market under its own name. A deployer uses a system under its own authority.

Reviewed 2 Oct 2026.

Was this page useful? Tell us what is missing or wrong ↗ · Report a correction