AI Acceptable Use Policy: Refunds, Disclosures And Contract Risks To Watch

AI is transforming SaaS, ecommerce and digital platforms, but it also brings new legal risks. If your business uses AI or offers AI-powered features to customers, your acceptable use policy (AUP) needs to do more than just restrict bad behavior. It must address how refunds work, what you disclose about AI, and which contract risks you need to manage. Many founders and operators overlook these issues, leading to customer disputes, chargebacks, or even FTC investigations.

Common mistakes include failing to tell customers when they are interacting with AI, copying generic terms that do not fit AI use cases, or not updating refund policies for AI-driven products. This guide explains what an AI acceptable use policy should cover, where US law sets the baseline, and what practical steps you can take to reduce risk. We include state-law caveats, real-world examples, checklists, and common pitfalls for SaaS, ecommerce and platform businesses.

What Is An AI Acceptable Use Policy?

An AI acceptable use policy (AUP) is a set of rules and disclosures that tells users how they can use your AI-powered features, what limits apply, and what happens if something goes wrong. It is usually part of your customer terms, SaaS agreement, or platform terms of service. For ecommerce, it may also appear in your website terms or product-specific policies.

Unlike a standard AUP, an AI AUP must address unique risks:

• AI-generated outputs may be inaccurate or unpredictable. Customers may expect more than your AI can deliver.
• AI may use or process sensitive data. This raises privacy and data security issues, especially under state laws.
• AI features may rely on third-party APIs or models. You need to disclose this and explain any limits.
• Refunds for AI-driven services may be limited. For example, if AI creates custom content, it may not be refundable.

For example, a SaaS tool that uses AI to summarize legal documents should tell users that the summaries are automated, may contain errors, and are not legal advice. An ecommerce site using AI chatbots should disclose that customers are interacting with AI, not a human, and clarify what happens if the chatbot gives incorrect information.

Without clear terms, customers may file chargebacks, leave negative reviews, or complain to regulators. A strong AI AUP helps set expectations, reduce disputes, and comply with US law.

Federal Rules: Disclosures, Refunds And Negative Option Risks

At the federal level, the Federal Trade Commission (FTC) is the main regulator for unfair or deceptive practices in advertising, disclosures, and refunds. Several FTC rules and guidance documents affect how you present your AI features and customer terms.

• Advertising and AI claims: The FTC requires that all marketing claims about AI be truthful and not misleading. If you overstate what your AI can do, or fail to disclose its limitations, you risk enforcement. For example, claiming "AI-powered legal advice" when your tool only summarizes documents could be deceptive.
• Negative option and auto-renewal rules: If your AI service is sold as a subscription, the FTC requires clear, upfront disclosures about recurring charges, how to cancel, and refund rights. The FTC's negative option guidance warns against hiding key terms or making cancellation difficult.
• Refund and cancellation policies: The FTC expects refund terms to be clear and conspicuous. If your AI product has unique refund limits (such as custom AI outputs), you must explain this before purchase. Burying refund terms in fine print or failing to honor stated policies can lead to regulatory action.
• Disclosures about AI use: If customers interact with AI (such as chatbots, recommendation engines, or automated decision tools), the FTC expects clear disclosure. Failing to tell customers they are dealing with AI, or that results are automated, can be seen as misleading.

For example, if your SaaS platform uses AI to generate investment recommendations, you must make it clear that the output is automated, may not be accurate, and is not a substitute for professional advice. If refunds are only available in limited cases due to the nature of AI, this must be stated clearly before purchase.

These federal rules apply nationwide, but state laws and industry-specific rules can add extra requirements.

State Laws: Auto-Renewal, Refunds And AI-Specific Rules

Many states have their own laws on auto-renewal, refunds, and disclosures, which can be stricter than federal rules. If you serve customers in multiple states, you need to consider the most restrictive requirements.

Auto-renewal laws: States like California, New York, and Vermont have detailed laws for auto-renewing subscriptions. These often require:

• Clear, conspicuous disclosure of renewal terms before purchase
• Affirmative consent to recurring charges (such as a checkbox)
• Renewal reminders before each renewal period
• Easy, online cancellation methods if signup was online

For example, California's Automatic Renewal Law (ARL) requires businesses to let customers cancel online if they signed up online, and to send a notice before a free trial converts to a paid subscription. If your AI SaaS or platform product uses auto-renewal, you must follow these state-specific requirements, not just the FTC rules.

Refund and return laws: Some states require specific disclosures about refund rights and may mandate minimum refund periods for digital goods. For example, California requires clear disclosure if a product is not refundable. If your AI product is not refundable, or refunds are limited due to the nature of AI outputs, you must state this clearly and comply with any state-specific requirements.

AI-specific rules: While most states do not yet have AI-specific consumer laws, some (like Illinois and Texas) have rules about biometric data, automated decision-making, or AI use in employment. If your AI tool processes personal data or makes decisions affecting users, you may need to provide extra disclosures or obtain consent under state law. For example, Illinois' Biometric Information Privacy Act (BIPA) requires written consent before collecting biometric data.

State law can also affect how you handle disputes, data privacy, and even how you notify users about changes to your AUP. If you operate nationwide, review your terms for compliance with both federal and the strictest applicable state rules.

Key Clauses To Include In Your AI Acceptable Use Policy

To reduce legal risk and set clear expectations, your AI acceptable use policy should address several key areas. Here is a practical checklist for SaaS, ecommerce, and platform businesses:

• Disclosure of AI use: Clearly state when and how AI is used in your product or service. For example, "Some features of this platform use artificial intelligence to generate responses or recommendations."
• Limitations and accuracy: Explain any known limitations of your AI, such as potential for errors, bias, or incomplete results. For example, "AI-generated results may not be 100 percent accurate and should not be relied on as the sole basis for decisions."
• Refund and cancellation terms: Specify when refunds are available, any limits due to the nature of AI outputs, and how customers can request a refund. For example, "Due to the automated nature of AI outputs, refunds are only available if the service fails to deliver as described."
• Prohibited uses: List unacceptable uses of your AI features, such as using the service for unlawful purposes, generating harmful or offensive content, or attempting to reverse-engineer the AI.
• User data and privacy: Explain how user data is used, stored, and protected, especially if AI processes personal information. Reference your privacy policy and any special rules for biometric or sensitive data.
• Third-party AI tools: If you use third-party AI providers or APIs, disclose this and explain any relevant terms or limitations. For example, "This service uses third-party AI models. Outputs may be subject to the terms of those providers."
• Support and dispute process: Describe how customers can get help if they have issues with AI features, and how disputes will be handled.

Here is a sample clause for disclosure and limitations:

This service uses artificial intelligence (AI) to generate responses, recommendations or content. AI-generated outputs may not be accurate, complete or appropriate for all situations. Users should exercise their own judgment and seek professional advice where necessary. The company does not guarantee the accuracy or reliability of AI-generated results.

Review your policy regularly as your AI features evolve, and update terms if you add new capabilities or change refund practices. If you are unsure whether your policy covers all necessary areas, consider seeking advice from a legal professional experienced in Software & IT or ecommerce matters.

Checklist for drafting or updating your AI AUP:

• Identify all AI-powered features in your product or service
• List any third-party AI providers or APIs you use
• Document known limitations and potential risks of your AI
• Draft clear disclosures about AI use and limitations
• Set refund and cancellation terms specific to AI outputs
• List prohibited uses and explain enforcement actions
• Reference your privacy policy and any special data rules
• Describe support and dispute resolution processes
• Review state-specific rules for auto-renewal and refunds
• Update your policy as AI features or laws change

For example, a SaaS founder launching an AI-powered analytics dashboard should review the dashboard's outputs for accuracy, clarify that insights are automated, and explain refund limits if the dashboard is wrong. An ecommerce operator adding an AI chatbot should disclose that the chatbot is AI, explain what it can and cannot do, and provide a clear process for customers to report errors or request refunds.

Common Mistakes And How To Avoid Them

Many startups and online businesses make avoidable mistakes when drafting or updating their AI acceptable use policy. Here are some of the most frequent issues and practical tips to avoid them:

• Copying generic terms: Using a standard SaaS or website terms template without adapting it for AI features can leave out critical disclosures or refund terms. Always tailor your policy to your actual AI use cases.
• Failing to disclose AI use: Customers may feel misled if they do not know they are interacting with AI. This can trigger complaints or refund demands, especially if the AI makes a mistake.
• Unclear refund policies: Not explaining when refunds are (or are not) available for AI-driven services can lead to chargebacks or regulatory scrutiny. Be specific about refund limits and the reasons for them.
• Ignoring state auto-renewal laws: Many founders only check federal rules, but state laws can be stricter. If you sell subscriptions, make sure your auto-renewal and cancellation terms meet state requirements.
• Not updating terms as AI evolves: AI features change quickly. Failing to update your policy when you add new capabilities or change how AI is used can create gaps in your legal protection.
• Overpromising AI capabilities: Marketing your AI as "fully accurate" or "guaranteed" can create legal risk if it does not perform as promised. Be realistic and clear about what your AI can and cannot do.
• Overlooking third-party AI terms: If you rely on external AI providers, you may be subject to their terms. Failing to pass these on to your users or disclose relevant limits can cause disputes.
• Missing privacy and data disclosures: If your AI processes personal or sensitive data, you may need special disclosures or consents under state law. For example, biometric data in Illinois or facial recognition in Texas.

To avoid these mistakes, review your AI acceptable use policy at least quarterly, and whenever you launch new features or change your business model. Get input from your product, legal, and customer support teams to ensure your terms match how your AI actually works. Consider a "red team" exercise where team members try to find gaps or ambiguities in your AUP from a user's perspective.

For example, a founder who adds an AI-powered resume screening tool should check whether the tool processes sensitive data, whether it could create bias, and whether state employment or privacy laws require extra disclosures. The founder should also clarify in the AUP that the tool is automated, may make mistakes, and is not a substitute for human review.

FAQs

Do I need to tell customers when they are using AI features?

Yes, the FTC expects businesses to disclose when customers are interacting with AI, especially if it affects their decisions or experience. Clear disclosure helps manage expectations and reduces the risk of complaints or regulatory action. Some states may also require specific disclosures for automated decision tools or biometric data.

What should I include in my refund policy for AI products?

Your refund policy should explain when refunds are available, any limits due to the nature of AI outputs, and how customers can request a refund. If AI results are not always accurate or are customized, state this clearly before purchase. Check state laws for any required minimum refund periods or disclosure requirements.

Are there special rules for AI subscriptions or auto-renewals?

Yes, both federal and state laws require clear, upfront disclosures for auto-renewing subscriptions, including renewal terms, cancellation methods, and refund rights. States like California, New York, and Vermont have strict requirements for online cancellation and renewal reminders. If you serve customers in these states, make sure your terms comply with the strictest rules.

What happens if my AI makes a mistake or gives a bad result?

You should explain the limitations of your AI in your acceptable use policy, and clarify that users should exercise their own judgment. If a mistake leads to a customer dispute, your clear disclosures and refund terms can help resolve the issue and reduce legal risk. For high-risk use cases, consider additional disclaimers or support processes.

Do I need to update my AI acceptable use policy as my product changes?

Yes. As you add new AI features or change how your AI works, you should review and update your AUP. Laws and customer expectations are evolving quickly, so regular updates help keep your business protected and your customers informed.

Key Takeaways

• An AI acceptable use policy should clearly disclose when and how AI is used, explain limitations, and set out refund and cancellation terms.
• Federal FTC rules require truthful advertising, clear disclosures and fair refund practices for AI-driven products and services.
• Many states have stricter laws on auto-renewal, refunds and disclosures, so check your terms against state requirements if you operate nationwide.
• Common mistakes include copying generic terms, failing to disclose AI use, missing privacy disclosures, and not updating policies as AI features change.
• Review and update your AI acceptable use policy regularly, and involve your legal, product and support teams in the process.

If you need help reviewing or updating your AI acceptable use policy, or want to check your terms for SaaS, ecommerce or platform compliance, contact our team at (888) 449-8437 or team@sprintlaw.com. Where legal services are required, they are delivered by licensed lawyers at trusted US law firms through the Sprintlaw platform.