Alex is a co-founder of Sprintlaw. Alex previously worked at a top-tier firm as a lawyer specialising in technology and media contracts, and founded a digital agency which he sold in 2015.
AI can make your website or platform feel faster, smarter and more helpful. Maybe you're using it to summarise articles, suggest products, generate reports, answer customer questions or give users a first draft of something they can edit.
That's exciting, but it also raises a very practical question: what should you tell users about the AI output they see?
An AI output disclaimer can help set expectations, explain limits and reduce risk. But it isn't a magic shield. In New Zealand, your disclaimer still needs to sit alongside your obligations under laws like the Fair Trading Act 1986, Consumer Guarantees Act 1993 and Privacy Act 2020.
Do You Need An AI Output Disclaimer On Your Website?
Yes, you should usually have an AI output disclaimer if your website or platform displays AI-generated content, recommendations, summaries, ratings, prompts, reports or answers to users.
The point of the disclaimer is to tell users, in plain language, that some content is generated or assisted by AI and may not always be complete, current or accurate. This matters because AI tools can produce confident-sounding answers that are wrong, out of date or missing key context.
You might need an AI disclaimer if your business uses AI to provide:
- product recommendations or comparison summaries
- health, wellness, finance, legal, HR or compliance information
- automated customer service replies
- draft documents, templates, captions, reports or emails
- search summaries or article summaries
- risk scores, suitability scores or eligibility guidance
The more a user might rely on the output to make a decision, the clearer your disclaimer should be.
For example, if you run a skincare eCommerce store and your website uses AI to suggest a routine based on a customer's answers, you should make it clear that the recommendation is general, may not suit every skin type, and doesn't replace advice from a qualified health professional where needed. You'd also want to make sure your website copy and advertising claims about the products are accurate and supported.
A disclaimer is only one piece of the puzzle. Your website terms of use, privacy policy, internal AI review process and customer support process all need to work together.
What Should An AI Output Disclaimer Say?
An AI output disclaimer should clearly explain what the AI does, what it doesn't do, and what users should do before relying on the output.
Good AI disclaimers are specific. A vague line like AI content may be inaccurate is better than saying nothing, but it often won't be enough if users are making meaningful decisions from the output.
Your disclaimer should usually cover:
- AI involvement: tell users that some content is generated, summarised, ranked or suggested using AI.
- Accuracy limits: explain that AI output can contain errors, omissions or outdated information.
- No substitute for professional advice: where relevant, say users should get advice from an appropriately qualified professional before acting on the output.
- User responsibility: ask users to check the output against their own circumstances before relying on it.
- Human review: explain whether your team reviews all, some or none of the AI output before it appears.
- Data limits: tell users not to enter confidential, sensitive or unnecessary personal information unless your platform is designed to handle it safely.
- Reporting errors: give users a simple way to flag wrong, unsafe or harmful output.
The wording should match how your tool actually works. If your team doesn't review every AI answer, don't say that it does. If the AI is trained only on your internal knowledge base, say that rather than implying it searches everything online.
This is where many templates fall short. They can give you a starting point, but your disclaimer works best when it's tailored to your business model, your AI tool and the level of risk in the output.
Can An AI Disclaimer Remove Your Legal Responsibility?
No, an AI disclaimer can't remove all of your legal responsibility, especially if your business makes misleading claims or provides services that don't meet New Zealand consumer law standards.
Under the Fair Trading Act 1986, you must not mislead or deceive customers, or make false or misleading representations about your goods or services. Calling something AI-generated doesn't let you publish inaccurate claims about prices, features, availability, results or suitability.
For some Fair Trading Act breaches, companies can face penalties of up to $600,000, and individuals can face penalties of up to $200,000. That doesn't mean every mistake will lead to a penalty, but it shows why AI output needs proper controls.
The Consumer Guarantees Act 1993 may also apply if you supply goods or services to consumers. For example, your services generally need to be provided with reasonable care and skill and be fit for purpose where the customer has made that purpose known. A disclaimer can't simply contract out of consumer guarantees for ordinary consumer transactions.
A Worked Example
Say Hana runs a New Zealand online fitness platform. Her website uses AI to generate personalised workout suggestions from a user's age, goals and equipment.
The AI gives one customer a high-intensity programme, even though the customer entered that they were recovering from a knee injury. Hana's disclaimer says the output is AI-generated and users should check suitability before starting. That helps, but it may not be enough if the website also markets the tool as safe personalised training for all injuries and no human has reviewed the outputs.
A better setup would include:
- clear warnings that the AI doesn't provide medical or physiotherapy advice
- prompts telling users with injuries to speak to a health professional
- filters that stop high-risk recommendations from being generated
- human review for higher-risk programmes
- a simple process to report unsafe recommendations
The disclaimer supports the risk controls. It doesn't replace them.
How Should You Handle Accuracy Warnings And Human Review?
You should be honest about the level of human review and use accuracy warnings that reflect the real risk of the output.
If your website says AI answers are checked by your team, users will reasonably expect that a human has reviewed the relevant output before they rely on it. If that's not true, the statement itself may create Fair Trading Act risk.
Think about your AI outputs in three practical risk levels:
Low-Risk Output
This might include AI-generated blog summaries, product descriptions that your team approves before publishing, or suggested captions for users to edit.
Your disclaimer can usually be lighter, but you should still say that the content may contain errors and users should check details before relying on it.
Medium-Risk Output
This might include product recommendations, comparison results, automated helpdesk answers, HR policy summaries or business compliance checklists.
You may need stronger warnings, human spot checks, a review process for flagged outputs and limits on what the AI can answer.
Higher-Risk Output
This might include health, financial, legal, employment, safety, insurance or eligibility recommendations.
For higher-risk output, a disclaimer alone is rarely enough. You should consider mandatory human review, clearer limits, escalation to qualified professionals, and stronger controls over the data the AI uses.
If your AI is giving employment-related guidance, keep it current. New Zealand employment law changes can affect the answer, such as changes to contractor status rules under the Employment Relations Act 2000, including the contractor gateway test changes introduced by the Employment Relations Amendment Act 2026. If your AI is using old content, your users may get the wrong answer.
What Privacy Issues Should Your AI Disclaimer Cover?
Your AI disclaimer should connect with your privacy approach if users can enter personal information into the tool.
New Zealand's Privacy Act 2020 applies when your business collects, uses, stores or discloses personal information. If your AI tool collects personal information, your users should know what you collect, why you collect it, who you share it with and how they can access or correct it.
Your AI disclaimer doesn't replace your privacy policy, but it can give users a timely warning at the point they use the AI feature.
For example, your disclaimer might tell users:
- not to enter sensitive personal information unless it's necessary for the service
- whether their prompts or uploaded files may be reviewed by your team
- whether AI providers or other service providers may process their information
- whether their inputs may be used to improve the tool
- how to request access to or correction of their personal information
You should also think carefully before asking users to upload information about other people. For example, if your platform summarises employment documents, customer complaints, medical notes or student records, you may be handling sensitive information even if your business isn't in a regulated industry.
If there's a privacy breach that's likely to cause serious harm, the Privacy Act 2020 may require you to notify the Privacy Commissioner and affected individuals as soon as practicable. Your internal process should cover who investigates the issue, who contacts your AI provider, what records you keep and who makes the notification decision.
What Should You Include In An AI Disclaimer Template?
Your AI disclaimer template should be treated as an outline, not a finished legal document. Have a lawyer or legal team tailor it before you publish it, especially if your AI output affects decisions about money, health, safety, employment or legal rights.
A practical outline might look like this:
- AI Use Statement: explain which parts of your website or platform use AI-generated or AI-assisted output.
- Purpose Of Output: describe what the output is designed to do, such as summarise information, suggest options or generate a draft.
- Accuracy Warning: say that AI output may be incomplete, inaccurate, outdated or unsuitable for a user's circumstances.
- No Professional Advice: where relevant, state that the output doesn't replace advice from a qualified professional.
- Human Review Position: say whether output is reviewed by a person, and if so, when and how.
- User Checks: tell users what they should verify before relying on the output.
- Privacy And Confidentiality Warning: explain what users should avoid entering and where to find your privacy information.
- Prohibited Use: list any uses you don't allow, such as illegal, harmful, discriminatory or unsafe uses.
- Error Reporting: give users a simple way to report inaccurate, harmful or inappropriate output, alongside any general disclaimer wording.
- Related Terms: connect the disclaimer to your website terms, platform terms, subscription terms, privacy policy and acceptable use rules.
Place the disclaimer where users will actually see it. For example, you might include a short notice next to the AI tool, with fuller wording in your website terms and conditions. If the AI output is high risk, hiding the warning in a footer won't be enough from a practical risk perspective.
You should also review your disclaimer whenever the tool changes. If you switch AI providers, add new data sources, introduce automated decisions or remove human review, your disclaimer and terms may need updating.
Key Takeaways
- An AI output disclaimer helps set user expectations, but it doesn't remove your obligations under New Zealand law.
- Your disclaimer should explain what the AI does, its accuracy limits, whether humans review the output and what users should check.
- The Fair Trading Act 1986 still applies if AI-generated content misleads customers or overstates what your product or service can do.
- If users enter personal information, your AI disclaimer should work together with your Privacy Act 2020 processes and privacy policy.
- Higher-risk AI output, such as health, financial, legal, employment or safety guidance, usually needs more than a disclaimer.
- Templates can be useful, but your wording should be tailored to your AI tool, customer journey and risk level.
If you'd like help with your AI output disclaimer, website terms and conditions, privacy wording or platform risk settings, you can reach us at 0800 002 184 or team@sprintlaw.co.nz for a free, no-obligations chat.
Drafted by Taylor, Sprintlaw's AI, and overseen by Alex Solo. How we write
Get your customer-facing terms right
What should your privacy and online terms cover?
If you collect customer data, sell online or run marketing campaigns, your public terms and privacy documents should match the real customer journey.








