Loading the generator. If this message stays, JavaScript is switched off in this browser, and the generator needs it.
Human Compute AI Policy Generator
Answer a few questions and get a finished AI policy for your business, ready to sign. It covers what your staff may put into AI tools, and the part most policies skip: what your vendors and customers may do with your data.
The three questions to ask every vendor
- Have you ever sold any of my data to AI companies?
- Do you have a policy that prevents that?
- From today, I want to opt out of that in the contract: my data is never sold to an AI company or used to train an AI model. Will you add that?
Get them as a ready-to-send email, with the contract wording
- Built in your browser, nothing sent anywhere
- No sign-up, no email needed
- Word, PDF and Markdown
An AI policy has to work in both directions
Almost every AI policy template covers one thing: what your people may paste into a chatbot. That matters, and this one covers it properly. But the live problem now runs the other way. The software you pay for, the processors that hold your records and the customers you trade with all sit on your data, and it is worth money to anyone training an AI model.
Inward
What your people put into AI
Approved and banned tools, a plain data classification, a never-paste list of what must never go into a public model, the training setting, exports and retention, contractors and offshore providers, and what to do when something goes wrong.
Outward
What others may do with your data
A position your business adopts on vendors and customers selling or training on your data, a clause you can paste into your contracts, the three questions to ask every vendor, and a note for your suppliers on why this protects them too.
Is your vendor allowed to train on your data? Is it allowed to sell it to an AI company? Even a record as small as "this customer bought from you ten times in 2026"? If your contracts do not say, the answer may be yes.
And it reaches past you. When a customer sells its accounting and messaging archive to an AI developer, it also sells every supplier's prices, volumes and reorder cadence. That is competitive intelligence about businesses that never agreed to anything. The policy includes a note your suppliers can read, and a clause both sides can sign.
The two pieces to send today
Both are in your policy already. They are here on their own so you can copy them straight into an email or a contract. They fill in with your company name as you type.
The three questions for every vendor
The contract clause
Want a second pair of eyes on it?
I'm Adam Ridley, and Human Compute is my AI consultancy in Winston-Salem, NC. Leave your email and I will reply, and you can send me your filled-in policy so I can tell you what I would change for a business like yours. This is optional. The policy above is yours either way.
This form sends only what you type into it, never your answers above. It is emailed to me and used only to reply to you. It is never sold, licensed or shared with anyone, and never used to train an AI model. You can ask me to delete it at any time.
Questions about the AI policy generator
Is the Human Compute AI Policy Generator free?
Yes. There is no sign-up and no email needed. Answer the questions, then download the policy as a Word file, save it as a PDF, or copy it.
Is the generated policy legal advice?
No. It is a starting template written in plain English, not legal advice. It gives you a complete, workable policy on day one. Have a lawyer read it before you rely on it, especially in healthcare, finance, legal or any business with regulated data.
Does anything I type leave my browser?
No. The policy is put together on your own device, and your answers are never sent to a server or stored. The review form sends only what you type into that form. This site has no analytics or advertising trackers today. If that ever changes, for example to show ads to people who have visited, the privacy policy will say so first, and those tools would see that you visited, never what you typed.
What does the AI policy cover?
Scope, including contractors and offshore providers. Approved and banned tools. A four-level data classification with a never-paste list: what must never go into a public AI model. Whether vendors may train on your inputs and how to switch that off. Exports, history and retention. Third parties who hold your data. Checking AI output. Incident reporting. Review dates. A signature block for approval and for every person it covers.
Why does it cover what vendors do with my data?
Because that is where most businesses are exposed now. Your invoices, orders, support tickets and messages are valuable to anyone training an AI model, and many contracts written before 2023 say nothing about selling or training on them. The policy sets out a clear position, gives you a clause to put in contracts, and gives you the three questions to ask every vendor.
Does it cover contractors and offshore providers?
Yes. They are named in scope, they sign the same acknowledgement, and there is a section on disclosing the AI tools they use on your data, where that data is processed, and deleting it when the work ends.
How often should an AI policy be reviewed?
At least once a year, and more often if you are strict. The generator sets once a year for permissive, every six months for balanced and every three months for locked down, plus a review any time a new tool is adopted, a vendor changes its terms or something goes wrong.
Can I get help putting it in place?
Yes. Use the review form above, or get in touch. Human Compute helps businesses work out where AI fits and then builds it.