Provenance
    DirectoryUse casesTrust pagesAI ActPricingDocs
    Sign inStart free
    Provenance

    AI Act Article 50 trust pages for AI SaaS — public evidence profiles, disclosure templates, and hosted buyer-ready trust pages on the Naburis platform.

    Product

    • Directory
    • Use cases
    • Trust pages
    • Pricing
    • Docs

    Company

    • About
    • Contact
    • Changelog
    • EU AI Act

    Developers

    • API
    • ProvenanceBot
    • Webhooks
    • LLMs.txt

    Legal

    • Terms
    • Privacy
    • DPA

    Powered by Provenance — Article 50 trust pages and public-source vendor evidence.

    © 2026 Naburis. All rights reserved.

    ← Docs

    EU AI Act Article 50 — implementation guide

    EU AI Act Article 50 — applicable 2026-08-02

    Transparency obligations for chatbots, AI-generated content, deep fakes, and public-interest content take effect 2 August 2026. Provenance ships templates and machine-readable provenance metadata aligned with the European Commission's December 2025 Code of Practice draft.

    See the templatesStart a draft

    Article 50 transparency obligations require two tracks at once for most AI products: a visible disclosure the end user perceives, and a machine-readable mark that downstream systems can detect. Provenance ships both — templates for the visible disclosure plus JSON-LD provenance metadata baked into every trust page export.

    Two-track requirement

    The European Commission's Code of Practice on marking and labelling of AI-generated content workstream describes a multilayered approach: visible disclosures combined with invisible or machine-readable techniques (metadata or watermarking) to improve resilience against removal or manipulation. Most Article 50 tracks expect both layers.

    The four tracks

    Each track addresses a different combination of AI system + deployer role. Pick the ones that apply, fork the templates, edit for your product, and publish.

    • Article 50(1)

      AI interaction notice

      Applies to: Providers of AI systems intended to interact directly with natural persons (chatbots, voice assistants, AI agents).

      Visible: Inform end users they are interacting with an AI system, unless this is obvious to a reasonably well-informed person from the circumstances and context of use.

      Machine-readable: Surface the AI system identifier in the response envelope (HTTP header, JSON-LD `dct:type`, or system message metadata) so accessibility tools and downstream agents can detect the AI source.

      Examples: In-app chat assistant that helps users navigate the product · Voice agent that handles inbound support calls · Email triage bot that responds without human review

    • Article 50(2)

      AI-generated content marking

      Applies to: Providers of AI systems generating or manipulating text, image, audio, or video content (synthetic media).

      Visible: Mark output as artificially generated or manipulated in a way the end user can perceive — visible label, audio cue, or watermark depending on modality.

      Machine-readable: Apply machine-readable provenance metadata (C2PA-compatible signing, EXIF tags, ID3 frames, or robust watermarking) so detection mechanisms and downstream platforms can identify the content as AI-generated.

      Examples: AI-generated marketing copy or product descriptions · AI-generated product photography for e-commerce · AI-generated audio for podcast production

    • Article 50(4) §1

      Deep-fake disclosure

      Applies to: Deployers of AI systems generating or manipulating content constituting a deep fake (image, audio, or video resembling real persons, objects, places, entities, or events).

      Visible: Disclose that the content has been artificially generated or manipulated in a clear and distinguishable way at the latest at the time of first interaction or exposure.

      Machine-readable: Combine the deep-fake disclosure with provenance signals — content credentials, signing chains, or watermarks — that survive re-encoding and distribution. The multilayer approach combines visible disclosure with invisible or machine-readable techniques.

      Examples: AI-generated likeness in an ad creative · Voice cloning for customer support · Synthetic talking-head explainer videos

    • Article 50(4) §2

      Public-interest content disclosure

      Applies to: Deployers of AI systems generating or manipulating text published with the purpose of informing the public on matters of public interest.

      Visible: Disclose that the text has been artificially generated or manipulated, except where the AI-generated content has undergone human review or editorial control and where a natural person or legal person holds editorial responsibility for the publication.

      Machine-readable: Annotate the publication with machine-readable provenance + editorial-responsibility metadata (e.g. `prov:wasGeneratedBy` + `prov:wasAttributedTo`) so search engines, fact-checkers, and aggregators can surface the disclosure alongside the content.

      Examples: AI-assisted news articles published by a media organization · AI-generated political analysis or commentary · AI-summarized legal or civic information

    Localized starter templates

    Markdown drafts for the visible-disclosure track. Customize before publishing in-app. Files also live under docs/ai_act_article_50_templates/ in the repo so they version with the rest of your trust copy.

    • English template page · Download Markdown
    • Deutsch template page · Download Markdown
    • Français template page · Download Markdown
    • Italiano template page · Download Markdown

    Machine-readable mark

    Provenance trust pages expose a JSON-LD export at /trust/[slug]/export/jsonld. The export carries the disclosure body, a content hash, a timestamp, and provenance metadata (prov:wasGeneratedBy, prov:wasAttributedTo) that downstream systems — search engines, fact-checkers, content-credential platforms — can ingest directly. Compatible with C2PA-style content credentials when paired with a signing pipeline.

    For products outside trust pages (chatbot interfaces, image generators), the JSON-LD payload also serves as a canonical reference your runtime can link to from response metadata, EXIF tags, or watermarks.

    Stay current

    The Commission's Article 50 code work is active through 2026. We update these templates against public Commission drafts and guidance as they land. Subscribe to the changelog or watch the Provenance directory profile for new template releases.