中文sales@inspirentech.co
Home/Insights/How to write an AI acceptable-use policy: a starter checklist
Insights · 7 min read

How to write an AI acceptable-use policy: a starter checklist

A practical checklist for writing an AI acceptable-use policy in your company: approved tools, data rules, verification, accountability and review.

An AI acceptable-use policy tells employees what they may and may not do with AI tools. It does not need to be long. A clear one-to-two-page policy, understood by staff, does more than a detailed document nobody reads. Use this checklist as a starting point.

1. Purpose and scope

  • State why the policy exists: to enable safe, productive AI use
  • Define who it covers: employees, contractors, interns
  • Define which tools it covers: approved assistants, AI features in other software, personal accounts

2. Approved tools

  • List the AI tools and account types approved for work
  • State clearly that personal or free consumer accounts are not to be used for company data
  • Name an owner who approves new tools

3. Data rules

  • Define categories: public, internal, confidential, personal data
  • Specify what may be entered into approved tools, and what never may
  • Reference your obligations under Singapore's Personal Data Protection Act and any client contracts

4. Accuracy and verification

  • Require staff to verify facts, figures and citations before use
  • Require human review for customer-facing, legal, financial, HR and regulated content
  • Prohibit presenting unverified AI output as fact

5. Accountability and disclosure

  • The employee remains responsible for work produced with AI assistance
  • Decide when AI use must be disclosed to clients or colleagues
  • Prohibit using AI to make final decisions about people without human review

6. Intellectual property and confidentiality

  • Do not input trade secrets, source code or client IP into unapproved tools
  • Clarify ownership expectations for AI-assisted work

7. Training and support

  • Require basic AI literacy training before access to approved tools
  • Name a contact for questions and incident reporting

8. Incident handling and review

  • Define what counts as an incident, such as confidential data entered in the wrong tool
  • Explain how to report it, without blame for honest mistakes
  • Review the policy every six to twelve months, because tools change quickly

Reference frameworks

Singapore publishes useful guidance, including the Model AI Governance Framework for Generative AI developed by IMDA and the AI Verify Foundation. You do not need to implement everything, but it is a good reference for areas such as accountability, data and incident reporting.

Want help turning this checklist into a policy your teams will actually follow? Our responsible AI and governance training includes a policy-drafting session.

Frequently asked questions

How long should an AI policy be?

One to two pages is ideal for most SMEs. Clear rules and examples matter more than length.

Do we need legal review?

Yes, have your legal counsel or data protection officer review the final policy, especially around personal data and contracts.

This article is general information for business readers, not legal, financial or technical advice. AI tools and plans change quickly; check current vendor documentation before making decisions.

Take it further

AI governance and responsible AI training

A policy only works if people understand it. Our AI governance training and responsible AI training for employees turn this checklist into practical habits, and include a session on drafting your AI policy for your company.

Keep exploring

Ready to make AI work for your team?

Tell us about your team and goals. We'll reply with a tailored training plan and quote.

Get a Quote