AI Transparency Notice

A note on scope. disclos.eu has a small AI footprint. We use AI to help run our own audit work and to help draft articles. We do not run a chatbot, we do not generate content for you, and we do not infer anything about you. The sections below say exactly where AI sits and where it does not.

1. AI features in this product

For each feature above, the model provider, intended purpose, and output type are described in the model card maintained by Disclos Research Practice.

2. When you are interacting with AI

You do not interact with an AI system on disclos.eu. There is no chatbot and no assistant.

• We run no conversational AI, so no interaction label is needed.
• We publish no AI-generated image, audio, or video to you.
• We run no deepfake or synthetic-media feature.
• The three tools at /tools/* are rule-based calculators. They apply fixed rules a person wrote. They are not AI systems under Article 3 and they make no inference about you.

3. Emotion recognition and biometric categorisation

We do not use emotion recognition. We do not use biometric categorisation. We collect no biometric data of any kind.

4. Public-interest text

We use a large language model to help draft articles published at /article/*. A human reviews every published article and holds editorial responsibility for it before it goes live. Where AI assisted a draft, we say so in the article. This human-review step is the basis on which the Article 50(4) disclosure carve-out applies to our articles.

5. General-purpose AI (GPAI) models we use

We use Anthropic's Claude as a general-purpose AI model inside our own audit engine. We are a downstream deployer of that model. We are not a GPAI provider. Anthropic's upstream obligations under Article 53 sit with Anthropic, and its published documentation is at https://www.anthropic.com. We do not send your personal data to the model, and the model is not trained on our inputs.

6. Your rights

• Right to know. You can ask us at any time how AI is used in anything we publish or deliver.
• Right to human review. Every deliverable we ship is already human-reviewed. If you believe an automated step affected you, you can ask for a human to look again.
• Right to object. Where any AI processing relies on legitimate interest, you can object via the contact below.

Data subject rights under the GDPR are unaffected by AI use and are described in our Privacy Policy.

7. Contact

Questions about AI use: gatis@disclos.eu Data protection: gatis@disclos.eu