Commission publishes guidelines on transparency obligations for providers and deployers of certain AI systems (Article 50)

Tag: S-2026-07-20-eu-ai-office-transparency-guidelines Type: article (European Commission press release / news item, with linked Guidelines, Q&A and fact page) Author(s): European Commission — Directorate-General CONNECT / European AI Office Date of source: 2026-07-20 Date ingested: 2026-07-30 Authority weight: high — primary Commission communication announcing official Guidelines interpreting Article 50 of Regulation (EU) 2024/1689; retrieved in full via WebFetch from digital-strategy.ec.europa.eu. Raw file: S-2026-07-20-eu-ai-office-transparency-guidelines.md. External URLs: news item https://digital-strategy.ec.europa.eu/en/news/commission-publishes-guidelines-transparency-obligations-providers-and-deployers-certain-ai-systems ; press release https://ec.europa.eu/commission/presscorner/detail/en/IP_26_1653 ; Guidelines library page https://digital-strategy.ec.europa.eu/en/library/guidelines-transparency-obligations-providers-and-deployers-ai-systems ; Article 50 Q&A https://digital-strategy.ec.europa.eu/en/faqs/transparency-obligations-under-article-50-ai-act

What it claims

On 20 July 2026 the European Commission published Guidelines to assist providers and deployers of AI systems in meeting the AI Act’s transparency obligations under Article 50, which “start to apply on 2 August 2026”. The stated purpose is to help people recognise when they are interacting with AI, or when content has been generated or altered by AI, “reducing the risk of deception and manipulation”. The Guidelines “clarify which providers and deployers must comply with the transparency obligations for interactive AI systems and the marking and labelling of AI-generated content” — i.e. they define the scope of the Article 50 obligations.

The core obligations restated by the Commission:

• Providers must design AI systems so that users are informed when they are directly interacting with an AI, and must add machine-readable marks to enable detection of AI-generated or manipulated content.
• Deployers must inform people when they are exposed to deep fakes; to AI-generated content on matters of public interest published without human review or editorial control; and to emotion-recognition or biometric-categorisation systems.

The news item positions the Guidelines alongside three companion instruments: the Guidelines on Transparency of AI-Generated Content, an Article 50 Questions & Answers, a Quick Facts fact page, and the (previously published, 10 June 2026) voluntary Code of Practice on Transparency of AI-Generated Content. The Guidelines library entry states they “define the scope of transparency obligations for providers and deployers of AI systems under Article 50 of the AI Act”. Page “Last update”: 20 July 2026.

Notable quotes

“Today, the European Commission published guidelines to assist providers and deployers of artificial intelligence (AI) systems in meeting the AI Act’s transparency obligations, which start to apply on 2 August 2026.” — Commission news item, 20 July 2026

“Transparency obligations will help people recognise when they are interacting with AI or when content has been generated or altered by AI, reducing the risk of deception and manipulation.” — Commission news item, 20 July 2026

“Deployers will also have to inform people when they are exposed to deep fakes, to AI-generated content on matters of public interests without human review or editorial control, and to emotion recognition or biometric categorisation systems.” — Commission news item, 20 July 2026

What’s speculative vs. asserted

• Asserted: the 20 July 2026 publication date; that the Guidelines interpret Article 50 and clarify its scope; that the Article 50 transparency obligations start to apply 2 August 2026; the specific provider duties (user-interaction notice; machine-readable marking of AI-generated/manipulated content) and deployer duties (deep-fake disclosure; unreviewed AI-generated public-interest content disclosure; emotion-recognition / biometric-categorisation disclosure); the existence of the companion Q&A, fact page and voluntary Code of Practice.
• Nature of the instrument (as framed): these are Guidelines — interpretive/soft-law clarifying scope, not a new binding obligation in themselves; the underlying binding duties are Article 50 of Regulation (EU) 2024/1689. The Code of Practice referenced is voluntary.
• Not retrieved this run: the full text of the Guidelines PDF (only the announcement news item and its metadata were fetched); the detailed worked examples, exemptions, and any de-minimis carve-outs the Guidelines set out.

Topics this feeds

• EU AI Act — advances the Article 50 transparency track from the 8 May 2026 draft guidelines / consultation and the 10 June 2026 voluntary content-marking Code to final published Commission Guidelines (20 July 2026) interpreting the scope of the obligations, 13 days before the 2 August 2026 applicability date.

Open questions raised

• What precise scope boundaries, exemptions and worked examples do the full Guidelines set (only summarised at scope level in the announcement)? The detailed PDF was not retrieved this run.
• How do the binding Article 50 Guidelines interact with the voluntary Code of Practice on Transparency of AI-Generated Content — is Code adherence treated as a presumption of compliance, or merely evidence of good faith?
• For financial-services deployers, how will “AI-generated content on matters of public interest without human review” be read against existing Consumer Duty / communications-with-customers obligations, given the 2 August 2026 date and the December 2026 transparency-solution deadline set by the Digital Omnibus?