What is the key point?

From 2 August 2026, Article 50 transparency obligations are live. A credible response is not a generic policy. It is a system-level evidence trail showing scope, role, risk, disclosure decisions, responsible owners, supplier assurance and proof that the controls were tested.

What changed on 2 August 2026?

The EU AI Act now applies more broadly, including Article 50 transparency obligations for certain interactive and generative AI systems. Providers and deployers have different duties. The high-risk requirements do not all apply yet: Regulation (EU) 2026/1744 confirms amended application dates of 2 December 2027 for Annex III systems and 2 August 2028 for Annex I product-embedded systems.

The minimum evidence pack

  1. AI system inventory. Purpose, owner, supplier, users, geography, data, affected people, model and version.
  2. EU scope and role assessment. Why the organisation is a provider, deployer, importer, distributor or out of scope for each system.
  3. Risk classification. Prohibited-practice check, high-risk screening, transparency-risk mapping and rationale.
  4. Article 50 evidence. Notice wording, screenshots, machine-readable marking specifications, test results and editorial-review records where relevant.
  5. AI-literacy records. Role-based training, attendance, content, competence checks and refresh dates.
  6. Governance controls. Named owners, approval thresholds, human oversight, monitoring, incidents, exceptions and review cadence.
  7. Supplier due diligence. Contractual roles, system documentation, limitations, data terms, change notifications and assurance evidence.

Three UK business scenarios

A SaaS vendor selling into the EU

Map whether you are the provider of an AI system or are integrating another provider's model. Retain the product scope decision, transparency design, supplier evidence, customer instructions and test records.

A UK employer using third-party AI

Document the use case, affected employees or candidates, your deployer role, supplier checks, human oversight and whether the system could fall within an Annex III high-risk category.

A marketing team using generative AI

Record tools and workflows, confirm who holds editorial responsibility, define human review, preserve approvals for public-interest content and check whether synthetic media or deepfake disclosures are required.

A practical 30-day response

Week 1: validate the inventory and identify every EU-linked use case. Week 2: complete role, risk and Article 50 decisions. Week 3: implement or repair disclosures, training and supplier gaps. Week 4: test the controls and assemble an evidence index that a buyer, auditor or authority can navigate.

Frequently asked questions

Did the 2 August 2026 EU AI Act deadline pass?

Yes. Article 50 transparency duties and several governance and enforcement provisions have applied since 2 August 2026. High-risk system requirements follow the amended 2027 and 2028 dates.

What should a UK business evidence first?

Start with an AI system inventory, EU scope and role assessment, risk classification, Article 50 decisions, proof of disclosures, AI-literacy records, supplier evidence and named control owners.

Does the EU AI Act apply to UK companies after Brexit?

It can apply where a UK organisation places an AI system or model on the EU market, or where the output produced by its AI system is used in the EU. The assessment depends on the specific system and role.

Is an AI policy enough to demonstrate readiness?

No. A policy helps, but buyers and regulators may need operational evidence such as inventories, assessments, notices, test records, approvals, incident logs, training records and supplier due diligence.

Recommended next step

Start with the free AI Act Exposure Check. For a structured evidence review, book an AI Act readiness conversation. You can also use the detailed Article 50 transparency checklist.

Official sources

Last updated 4 August 2026. This article provides general information, not legal advice.