Corporate AI usage policy: 2026 template and legal obligations (GDPR, AI Act, security)
In brief
- The AI use charter governs uses, protects data and rights, and formalises the obligation to control AI (art. 4 AI Act). Integrated into the internal regulations or distributed with acknowledgement of receipt, it becomes binding on employees.
What an AI charter should contain
- Purpose and scope (who, which tools, which uses).
- Principles: human oversight, transparency, data minimisation.
- List of authorised tools and procedure for requesting new ones.
- Rules on data (what must never be entered into an AI).
- Output verification (hallucinations) and user responsibility.
- Confidentiality, content ownership, transparency towards third parties.
- Sanctions for breaches and contact point.
Permitted / prohibited uses: the framework table
| Category | Permitted (with precautions) | Prohibited / restricted |
|---|---|---|
| Data | Non-identifying, anonymised content | Personal or confidential data in a public tool |
| Decision | Aid, draft, summary | Automated decision with significant effect without human involvement |
| Content | Drafting, rephrasing | Distribution without verification or required disclosure |
Integrated into the internal regulations or distributed with acknowledgement of receipt, the charter becomes binding. If a tool affects working conditions, inform and consult the works council.
Take action
Train your teams and deploy a compliant AI charter (GDPR, AI Act art. 4 and 50) with Data Comply One.
Discover the AI Act software →Test my compliance for free