Labeled, Marked, Detectable: The EU’s AI Transparency Rules Land 2 August
July 2026
On 2 August, the EU AI Act’s transparency rules go live, and they overturn a quiet assumption behind a great deal of corporate content: that no one needs to be told when a machine did the writing, the drawing or the speaking. Article 50 applies across the bloc, and its reach extends to any organization whose systems or synthetic content touch the European market.1
What the rules actually require
Two duties sit at the center. Providers of generative AI systems must ensure their outputs — text, image, audio and video — are marked in a machine-readable format and detectable as artificially generated.1 Deployers, meaning the businesses using those tools, must tell people when they are interacting with an AI system, such as a chatbot, and must disclose AI-generated content in defined circumstances.1 Non-compliance can draw fines of up to €15 million or 3% of worldwide annual turnover, whichever is higher.2
The distinction that matters most for communicators
For marketing and corporate-affairs leaders, the practical line runs between text and synthetic media. AI-assisted text carries a disclosure duty only when it is published to inform the public on matters of public interest — and even then, the obligation falls away where the content has had genuine human review and a named person holds editorial responsibility.3 Ordinary brand and marketing copy, edited by a real person, generally needs no label. Synthetic audio and visuals are treated more strictly: if you use AI to create or alter images, audio or video that resemble real people, places or events and could appear authentic — a “deepfake” in the Act’s broad sense — you must label it, even in marketing, and even with an editor in the loop.1,3 The human-oversight carve-out covers text; it does not cover synthetic audiovisual content.
Implication: Read the line as “genuine editorial control” for text and “label it” for anything photorealistic. A nominal skim over an auto-published draft will not qualify for the exemption.
The August date survived the wider reset
The transparency deadline held even as Brussels softened the rest of the rulebook. On 29 June, the Council gave final approval to a simplification package that pushed high-risk deadlines to December 2027 and granted a narrow four-month reprieve — to 2 December — for marking content from systems already on the market; the headline transparency duties were left intact.4,5 A week later, on 7 July, the Commission published an EU Action Plan on Cybersecurity and Artificial Intelligence, signaling that the resilience of the most advanced models is now a security question, not only a compliance one.6
Implication: Do not read the 2027 extension as breathing room. If your products speak to customers or generate content, 2 August is the operative date.
What should leaders do now?
Three moves cover most of the exposure. Map where AI touches your published content, and separate text from synthetic media. Put a named editor and a real review step behind anything you intend to rely on the human-oversight exemption to cover. And build labeling and machine-readable marking into your image, audio and video pipelines now — retrofitting after publication is far harder than designing it in.
How Confident Strategy Group can help
The organizations that will move through this calmly are the ones treating transparency as an editorial governance question, not a legal afterthought. Confident Strategy Group helps leaders turn Article 50 into working practice — developing company policies, editorial responsibility protocols, and labeling standards that determine what gets disclosed, who signs off, and how AI-generated content is marked before it is published. And because the rulebook is still moving, our Confident ForeFront™ service keeps you ahead of it: a strategic radar that continuously scans the policy landscape, surfaces the shifts that matter to your business, and turns them into decisions before they become deadlines. If your content reaches European audiences, this is the moment to move from scramble to system.
Disclaimer: This briefing is provided by Confident Strategy Group for general information and strategic insight only and reflects the position as at July 2026. It does not constitute legal advice; organizations should consult qualified counsel on how the EU AI Act applies to their specific circumstances.
References
- EU AI Act, Article 50 — Transparency Obligations for Providers and Deployers of Certain AI Systems (Regulation (EU) 2024/1689), applicable 2 August 2026. Available at: https://artificialintelligenceact.eu/article/50/
- EU AI Act, Article 99 — Penalties. Available at: https://artificialintelligenceact.eu/article/99/
- Lausen. “Section 50(4) of the AI Act: what organizations must label as AI content from August 2026” explains the Article 50(4) human-review / editorial-responsibility exemption for text. Available at: https://lausen.com/en/section-504-of-the-ai-act-what-organisations-must-label-as-ai-content-from-august-2026/
- Council of the EU. “Artificial Intelligence: Council gives final green light to simplify and streamline rules.” 29 June 2026. Available at: https://www.consilium.europa.eu/en/press/press-releases/2026/06/29/artificial-intelligence-council-gives-final-green-light-to-simplify-and-streamline-rules/
- Sidley (Data Matters). “EU Lawmakers Reach Provisional Agreement to Delay Key EU AI Act Obligations.” 22 June 2026 — Article 50 duties remain on the 2 August 2026 timeline, with a four-month reprieve (to 2 December 2026) for marking content from systems already on the market. Available at: https://datamatters.sidley.com/2026/06/22/eu-lawmakers-reach-provisional-agreement-to-delay-key-eu-ai-act-obligations/
- European Commission. “Commission presents EU Action Plan on Cybersecurity and Artificial Intelligence.” 7 July 2026. Available at: https://digital-strategy.ec.europa.eu/en/news/commission-presents-eu-action-plan-cybersecurity-and-artificial-intelligence
- European Commission. “Code of Practice on marking and labeling of AI-generated content” (final, published 10 June 2026). Available at: https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content
Our Work >