Generative AI usage charter

This template is indicative. Adapt it to your organisation and have it reviewed by your legal adviser, in particular with regard to Algerian Law No. 18-07 of 10 June 2018 on the protection of personal data.

Generative AI usage charter

Template to adapt · [Company name]

1. Purpose

This charter sets the rules for using generative AI tools (chat assistants, image generators, translation or analysis tools) by the staff of [Company name]. It lets us benefit from these tools while protecting the company, its clients and its employees.

2. Principles

AI assists; humans decide and answer for the result.
All content is checked before it is used or shared.
Company and third-party data are protected.
Use of AI is disclosed where it matters.
The law, copyright and the rights of others are respected.

3. Which data may go into an AI tool?

Category	Examples	Rule
Public data	Press releases, already published texts, general documentation	Allowed
Internal, non-sensitive data	Document templates, generic procedures, anonymised texts	Allowed once anonymised, with a tool from Annex A
Sensitive data	Personal data (identity, salaries, health), trade secrets, contracts, client data, unpublished financial information, passwords	Never entered in an AI tool

4. Authorised tools

Only the tools listed in Annex A are authorised for professional use. Any new tool is submitted to the AI officer before use. A personal account is never used to process company data.

5. Checking before sharing

Read the whole text produced: AI can be wrong and make things up.
Check figures, names, dates, references and quotations.
Have any external or binding document validated by your manager.
Keep important prompts so you can explain the result.

6. Generated content

State that content was AI-generated when it is shared externally and substantially generated.
Never present a generated image as real.
Respect copyright and image rights.
Never reproduce the voice or face of a real person without written consent.

7. Uses

Allowed	With manager approval	Prohibited
Writing, rewording, translation, summaries of non-sensitive documents, brainstorming, outlines, tracking tables	Analysis of internal data, external communication, legal, financial or HR content, content published in the company’s name	Sole decisions about a person (hiring, sanction, appraisal) without human review; entering sensitive data; use to deceive; circumventing the charter

8. Roles

Management: approves the charter and arbitrates.
AI officer: keeps the list of tools, answers questions, collects incidents.
Managers: ensure application within their team.
Employees: apply, check, report.

9. Incidents

If sensitive data is entered or erroneous content is shared, inform the AI officer within 24 hours. The officer assesses, corrects and informs management.

10. Training and review

Every new employee is briefed on the charter. It is reviewed at least every six months.

Annex A · Authorised tools

Tool	Use	Authorised data	Owner

Annex B · The 7 questions before you send

Did I give the tool data that must stay confidential?
Did I read the entire text?
Are the figures, names and dates correct?
Do the sources cited really exist?
Are the tone and register right for the reader?
Should I disclose that AI took part?
Does my manager need to validate this document?

Read and approved · Name, position, date and signature:

Other languages: Français · English · العربية

Generative AI usage charter