EU AI Act

AI Act transparency disclosure.

Article 50 of Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 — the AI Act.

1. AI system disclosure

Jarvis is an AI system. Outputs you receive — written copy, social posts, customer replies, draft contracts, recommendations — are produced or assisted by general-purpose AI models, including Anthropic Claude and OpenAI GPT. Where required by Article 50, generated content is labelled as such.

2. Human oversight

Every Jarvis deployment runs in one of three control modes: Approval (every action signed off by you), Automation (you set policies, the system acts within them), or Hybrid (sensitive items approved, the rest automated). You can switch modes at any time.

3. Data and training

Your business data — customer records, brand voice, pricing decisions — is never used to train shared third-party models. Cross-client learning, where it occurs, is restricted to anonymised, aggregated patterns (for example: 'effective subject-line patterns in Czech retail').

4. Right to information

Customers interacting with Jarvis-generated content (chat replies, email responses) are informed when interacting with an AI system, in line with Article 50(1).

5. Risk classification

Jarvis is classified as a limited-risk AI system under the Act. We do not perform activities listed under Annex III as high-risk. Should our scope expand into high-risk areas, conformity assessments will be conducted before deployment.

6. Conformity timeline

EU AI Act provisions on general-purpose AI take effect 2 August 2026. Jarvis is conformity-ready by that date and continuously audited thereafter.

7. Contact for AI Act matters

Questions about our compliance, requests for technical documentation, or notifications under Article 50 should be sent to compliance@jarvis17.eu.