Free AI Policy Template for Small Business (US and EU)
A free one-page AI policy template for small businesses, with an EU add-on for the AI Act as amended in 2026, US rules, and a tool register.

A one-page AI policy is the cheapest protection a small business can put in place. Below is a free template to copy and adapt, a short EU add-on that reflects the AI Act as amended in July 2026, and a tool register to fill in. It takes about an hour.
This is a practical starting point, not legal advice. If AI touches hiring, credit, health or children in your business, have a lawyer read your version.
What the rules ask of a small business, in plain terms
In the EU
- Using ChatGPT at work makes you a “deployer”. The AI Act’s definition covers any business using an AI system under its authority, and the Commission’s own AI literacy Q&A uses staff using ChatGPT as its example.
- AI literacy is now a duty to support, not to certify. Since the AI Omnibus amendment took effect on July 27, 2026, Article 4 says providers and deployers “shall take measures to support the development of AI literacy”, and states that it does not require any specific level of literacy of any individual. No certificate is needed. A short session and a record of it is a sensible way to show you took measures.
- Some AI content must be labeled, since August 2, 2026. Under Article 50, deployers must disclose deepfakes (images, audio or video that could pass as real) and AI-generated text published to inform the public on matters of public interest, unless a person reviewed it and someone holds editorial responsibility. People must also be told when emotion recognition or biometric categorization is used on them.
- Fines are real but scaled. Breaches of Article 50 can reach EUR 15 million or 3% of turnover, and for small and medium businesses the lower of the two applies. Article 99 sets no fine tier for Article 4.
- GDPR still applies to every prompt. If staff paste customer or employee personal data into an AI tool, you need a lawful basis under the GDPR and, where the vendor processes that data for you, a data processing agreement under Article 28. Business plans usually offer one; free personal accounts usually do not.
In the US
- There is no single federal AI law for private businesses, but the FTC has said there is no AI exemption from the laws on the books. Its reviews rule also covers AI-written fake reviews.
- Hiring is where state rules bite first. New York City’s Local Law 144 has regulated automated hiring tools since July 5, 2023. Illinois limits AI in employment decisions under Public Act 103-0804, in effect since January 1, 2026. California’s Civil Rights Council rules on automated decision systems in employment took effect October 1, 2025.
- More is coming in 2027. Colorado replaced its original AI Act with SB26-189, with duties starting January 1, 2027, and California’s privacy agency rules on automated decision-making apply from January 1, 2027.
- A voluntary framework exists. The NIST AI Risk Management Framework is free and intended for voluntary use. Larger customers sometimes ask suppliers about it.
- Purely AI-made work may not be yours to protect. The US Copyright Office’s 2025 report says material generated entirely by AI is not copyrightable.
The template: copy, then fill in the brackets
[Business name] AI use policy (version 1, [date])
- Why we have this policy. We use AI tools to save time, and we stay responsible for everything they help produce.
- Approved tools. Only the tools in our tool register may be used for work, on business accounts. Personal AI accounts are not used for business data.
- Never put into an AI tool: passwords, API keys or login codes; bank or card numbers; health information; anything a client marked confidential; and personal data about customers or staff unless the tool is listed in the register as approved for it.
- A person checks before anything leaves the business. Facts, figures, names, prices and quotes in AI output are checked against a source before they reach a customer, a client or the public.
- When we label AI content. We label images, audio or video that could pass as real, and any chatbot tells people they are talking to AI. Content we publish to inform the public is either reviewed by a named person or labeled as AI-generated.
- No AI-only decisions about people. Hiring, firing, pay, credit and similar decisions are made by a person. AI may help prepare, never decide.
- Ownership. We keep our drafts and edits for important work, because work made by AI alone may not be protected by copyright.
- Learning. Everyone who uses AI tools gets a short introduction to what they do well, where they fail, and this policy. We note the date.
- Problems. If an AI tool leaks data, produces something harmful or makes a costly mistake, tell [name] the same day.
- Owner and review. [Name] owns this policy and reviews it every six months, or sooner if a tool or law changes.
EU add-on: three extra clauses
- AI literacy record. We keep a simple log of the measures we take to support AI literacy: date, who attended, what was covered. (AI Act Article 4, as amended in 2026.)
- Labeling under Article 50. Deepfakes are always disclosed. AI-generated text we publish on matters of public interest is disclosed unless a named person reviewed it and takes editorial responsibility.
- Personal data. Personal data goes only into tools covered by a data processing agreement, listed in the register, and only what the task needs.
The tool register
| Tool and plan | Who uses it | Data allowed |
|---|---|---|
| [Example: ChatGPT Business] | [Names or roles] | [Internal documents; no customer personal data] |
| [Example: Microsoft 365 Copilot Chat] | [Everyone] | [Emails and files already in our Microsoft account] |
| [Add a row per tool] |
Business plans of the main AI tools do not train on your data by default, while free and personal plans can unless you opt out. The AI plan price comparison lists which is which. For rolling a policy out in a week, the five-day governance plan fits alongside this template, and the EU AI Act transparency rules explain the labeling duties in more depth. More on rules and risks is on the Ethical AI topic page.
One step for today
Fill in the tool register before anything else. Listing which AI tools the business actually uses, and on which accounts, usually shows the one gap that matters most.
Rules checked on November 13, 2026, against official EU, US federal and state sources. This template is general information, not legal advice; laws differ by country and state and keep changing.
Sources
- AI Act Article 4, Article 50, Article 3 and Article 99 (European Commission AI Act Service Desk)
- AI Omnibus enters into force and AI literacy questions and answers (European Commission)
- Quick facts: transparency rules for AI systems (European Commission)
- General Data Protection Regulation and EDPB Guidelines 07/2020 on controllers and processors
- FTC Operation AI Comply, FTC on artificial intelligence and FTC reviews rule Q&A
- NIST AI Risk Management Framework
- NYC automated employment decision tools, Illinois Public Act 103-0804, Colorado SB26-189, California Civil Rights Council and California Privacy Protection Agency
- US Copyright Office, Copyright and Artificial Intelligence, Part 2
In this article

This Week in AI for Small Business: Claim Your Profile
AI news for small business, week of November 13, 2026: ChatGPT business profiles, Claude Sheets, stolen API keys and two December 11 deadlines.
Looking for something else?