Sustained FuturesChangemakers Collective
Sign in

AI policy starter kit

Write your AI policy in 10 minutes.

Answer 9 short questions and the policy builds on the right as you go: where you use AI, how you keep its energy and water footprint down, what a person always checks, and what you won't use it for. Copy it, download it, or save it as a PDF to sign.

Step 1

Who is the policy for?

Step 2

What do you use AI for?

Step 3

Which tools?

Step 4

What sensitive information do you handle?

Tick any that apply. Each adds a rule.

Step 5

Keeping the footprint down

All on to start. Untick any you won't commit to.

Step 6

What does a person always check?

Step 7

What won't you use AI for?

Step 8

When do you say AI helped?

Step 9

How often will you review it?

Sustained Futures · Changemakers Collective

My AI policy

Version 1, 4 November 2026

Why this policy exists

I use AI to do better work, faster, without losing sight of the people affected by it or the energy and water it takes to run. This policy sets out how I use it, what I won't do, and how I keep the footprint down.

Where I use AI

I use AI for everyday tasks where it saves time.

Keeping the footprint down

Cloud AI runs in data centres that use electricity and water. I treat that as a real cost.

  • I use AI where it saves real time or improves the work. A search, a template or a quick note to a colleague comes first when that does the job.
  • I use smaller, faster models for simple tasks and keep the largest models and long reasoning modes for work that needs them.
  • I give the context, the audience and the format up front, so I get a usable answer in fewer attempts.
  • Generating images, audio and video uses far more computing than text, so I only do it when the piece needs it, and I reuse what I've already made.
  • For simple or sensitive tasks, I use a small model that runs on my own computer where that works well enough.
  • When I choose between tools, I favour providers that publish their energy, water and emissions figures and their renewable energy commitments.
  • I cancel AI tools and automations I no longer use, and I don't leave agents running on a schedule that nobody reads.

Data and confidentiality

  • I check a tool's privacy settings before using it for work, and switch off training on my chats where the tool allows it.

A person checks the work

AI gets things wrong, confidently. A person reads and checks these before they go anywhere:

  • anything published or posted under my name
  • anything sent to clients, funders or partners
  • every number, claim and source, checked against the original

The person who signs the work off is responsible for it, whoever drafted it.

What I won't use AI for

I won't:

  • let AI make a decision about a person on its own
  • make images, audio or video of a real person without their consent
  • use AI to produce climate or impact claims I can't evidence
  • paste passwords, API keys or access codes into an AI tool
  • present AI output as research or analysis I did myself

Being open about AI

I say when AI shaped the substance of a piece of work: the analysis, the structure or the findings. I don't flag spellchecks and small edits.

Review

I review this policy every 6 months, and whenever my tools or my work change. Next review: 4 May 2027.

Signed

Name: ____________________

Signature: ____________________

Date: 4 November 2026

Made with the Changemakers Collective AI policy starter kit, Sustained Futures. General guidance, not legal advice. If your employer has its own AI policy, theirs comes first.

Looking for tools that fit your policy? See Responsible AI tools. General guidance, not legal advice. If your employer has its own AI policy, theirs comes first. Your answers stay in this browser; nothing is sent to Sustained Futures.