Black Friday10% off any legal service

Is It Legal To Use AI-Generated Customer Reviews?

AI can make marketing much faster, especially when you're trying to turn messy customer feedback into polished website copy. It's also tempting to ask an AI tool to create a few glowing testimonials so your product page doesn't look empty.

That's where the legal risk starts. In New Zealand, reviews and testimonials are advertising claims. If they're made up, exaggerated or presented in a way that gives customers the wrong impression, they can breach the Fair Trading Act 1986.

The good news is that you can still use AI in your marketing. The key is to use it as a drafting or summarising tool, not as a way to invent customer experiences that never happened.

Can You Use AI-Generated Customer Reviews In NZ?

No, you shouldn't publish AI-generated customer reviews if they look like genuine reviews from real customers but aren't based on real customer experiences.

Under the Fair Trading Act 1986, your business must not engage in misleading or deceptive conduct, or make false or misleading representations about goods or services. A customer review is powerful because readers assume it's independent feedback from someone who has actually dealt with your business.

If an AI tool writes a review such as "Best accounting software I've ever used - saved me 10 hours a week", but no customer said that, it's likely to be misleading. It doesn't matter that a human didn't personally write the fake review. Your business is still responsible for the advertising you publish.

The Commerce Commission can take action for breaches of the Fair Trading Act. For certain breaches, companies can face penalties of up to $600,000 per offence, and individuals can face penalties of up to $200,000 per offence. There can also be practical damage, such as refund requests, platform takedowns and loss of customer trust.

A safer way to use AI is to start with real feedback and keep the meaning accurate. For example, AI can help you:

  • turn a long review into a shorter quote, as long as the key message doesn't change
  • group common themes from genuine reviews, such as fast delivery or helpful support
  • draft a plain-English summary that you then check against the source reviews
  • remove personal details from feedback before publishing, where needed

Think of AI as your editor, not your customer.

When Does An AI Review Become Misleading?

An AI review becomes misleading when it creates an impression that isn't supported by real customer feedback or real evidence.

This can happen even if each sentence sounds fairly harmless on its own. The legal question is often about the overall impression your website, ad or product page gives to an ordinary customer.

Common Risky Examples

  • Invented testimonials: AI creates a quote from "Sarah, Auckland" even though Sarah doesn't exist.
  • Inflated results: a real customer said your course was useful, but the AI version says it "doubled my revenue in 30 days".
  • Fake volume: your website says "loved by thousands of customers" when you've only had 37 paying customers.
  • Cherry-picked summaries: your AI summary says "customers consistently praise our delivery speed" when many reviews mention late delivery.
  • Hidden incentives: you publish glowing reviews from customers who received a discount or free product, without making that clear where it would affect the customer's view.

Say Mike runs a cleaning business in Wellington. He has 12 genuine Google reviews. Eight mention good communication, three are neutral and one complains about missed spots. If Mike asks AI to write a website banner saying "Every customer says we're the most thorough cleaners in Wellington", that claim is a problem. It suggests unanimous praise and a market-leading comparison, neither of which he can prove.

A better version would be: "Customers often mention our clear communication and friendly service." Mike should still check that this is a fair summary of the actual reviews before publishing it.

The same rules apply across your website copy, social media ads, email campaigns, product pages and pitch decks. If the review or claim helps persuade a customer to buy, treat it as advertising.

Can AI Summarise Real Customer Reviews?

Yes, AI can summarise real customer reviews, as long as the summary is accurate, balanced and based on reviews you can point back to.

This is often the most useful and lowest-risk way to use AI in review marketing. Customers don't always write in neat, quotable sentences. AI can help you identify themes and convert feedback into readable marketing copy.

The key isn't to let the tool over-polish the feedback. AI tools are designed to be persuasive, so they can add certainty, emotion or results that weren't in the original review.

A Practical Review Summary Process

  1. Collect the source reviews: keep the original review text, date, platform and customer name or identifier where available.
  2. Ask AI for themes, not hype: use prompts such as "summarise the most common themes without adding claims that aren't in the reviews".
  3. Check negative and neutral feedback: don't let the summary imply every customer had the same positive experience if they didn't.
  4. Match claims to evidence: if the summary says "fast delivery", check that several reviews actually mention delivery speed.
  5. Keep a record: save the final wording and the reviews it was based on, so you can justify the claim later.

For example, imagine you run an online skincare store with 80 verified reviews for a moisturiser. If 52 reviews mention that the product feels lightweight, AI could help you write: "Many customers describe the moisturiser as lightweight and easy to apply."

But if only three reviews mention reduced redness, you should be careful with a headline like "Customers love it for calming redness". That may overstate how common the experience is. If the product claim also edges into health or therapeutic territory, you'll need to be even more careful about the rules that apply to those claims.

If you're using review platforms or marketplace tools, check their platform terms and rules too. Some platforms ban manipulated, incentivised or AI-generated reviews even where the legal position depends on the details.

What Counts As Proof For Customer Testimonials?

You need enough evidence to show that each testimonial is real, accurate and not taken out of context.

The Fair Trading Act doesn't use a special rule just because a claim appears in a testimonial. If you publish the testimonial, your business is making or adopting that claim. You should be able to back it up.

Good records can make a big difference if a customer, competitor, platform or regulator asks questions later.

Evidence To Keep

  • the original review or feedback, before editing
  • the customer's consent to use their words, name, photo or business name, where you have obtained it
  • the date the review was given and the product or service it relates to
  • proof that the customer actually bought or used the product or service, where this is relevant
  • details of any incentive, discount, gift or affiliate arrangement
  • the final published wording and where it appeared

If a testimonial includes a result, keep evidence for that result. For example, "We cut admin time by 30%" is more than a feel-good comment. It's a measurable claim. You'd want to know how the customer measured it, what period it covered and whether the result is typical.

Privacy also matters. If a review identifies a person, or you use a customer's name, photo, voice or other personal information, the Privacy Act 2020 can apply. You should be clear about what you're collecting, how you'll use it and whether the customer can ask you to remove or update it.

If you materially edit a review, don't make it look like an exact quote unless it still reflects what the customer actually said. A short quote is fine, but changing the meaning isn't.

How Should You Review AI Advertising Claims?

You should review AI advertising claims before they go live, because your business is responsible for the final wording even if AI helped create it.

This applies to review summaries, star-rating blurbs, comparison claims, chatbot-generated landing page copy and social ads. The safest habit is to build a simple sign-off process before anyone in your team publishes AI-assisted content.

AI Marketing Claims Checklist

  • Is the claim based on real evidence? If not, remove it or rewrite it.
  • Would an ordinary customer read it as a promise? If yes, make sure you can deliver on that promise.
  • Does it exaggerate frequency? Words like "everyone", "always", "guaranteed" and "number one" need strong proof.
  • Does it hide important context? For example, a result may only apply to premium customers or a particular use case.
  • Is there a comparison? Claims that you're cheaper, faster or better than competitors need current evidence.
  • Does the claim involve health, finance or safety? These areas often need extra checks because the consequences of misleading customers can be higher.
  • Have you checked the final version, not just the prompt? AI can add new claims during editing.

You should also consider the Advertising Standards Authority codes, especially if you're running public advertising campaigns. While the ASA process is separate from the Commerce Commission, complaints can still create cost, delays and reputational issues.

For higher-risk campaigns, it can be worth getting an advertising claims review before launch. This is especially true if you're using testimonials to support performance claims, running paid ads at scale or entering a market where competitors are likely to scrutinise your wording.

This doesn't have to slow your marketing team down. A clear internal checklist and a quick review of approved claim types can help your team move quickly while staying within sensible boundaries.

What Should Your Website Terms Say About Reviews?

Your website terms should explain how reviews may be submitted, moderated, displayed and removed, especially if customers can post reviews directly on your site.

Terms won't make a misleading review lawful, and they won't fix fake testimonials. But they can set clear rules for your customers and give your business a practical process for managing review content.

If you publish customer reviews, your customer terms and internal review policy should work together. Templates can be a useful starting point, but review terms work best when they're tailored to your website, sales process, refund policy and moderation practices.

Review Terms Outline

  1. Who can submit reviews: explain whether reviews are limited to verified customers or open to all users.
  2. Review standards: say that reviews must be honest, relevant and based on a genuine experience.
  3. Prohibited content: cover fake reviews, offensive content, confidential information, competitor spam and unlawful claims.
  4. Moderation rights: explain when you may edit, refuse, remove or hide a review.
  5. Use of reviews in marketing: state whether you may quote, summarise or display reviews in ads, emails or product pages.
  6. Personal information: explain how names, photos and other personal information may be used, consistent with your privacy practices.
  7. Incentives: set rules for discounts, gifts or rewards in exchange for reviews, including disclosure requirements.
  8. AI use: explain whether AI may be used to summarise reviews, and commit to not changing the substance of the feedback.
  9. Contact process: tell customers how to raise concerns about a review or request an update.

A legal professional should tailor this before you use it. For example, a SaaS business collecting in-app reviews will need different wording from an ecommerce store publishing product reviews or an agency using client case studies.

You should also make sure your website terms line up with your privacy policy, social media processes and any agreements with influencers, affiliates or customers who provide case studies.

Key Takeaways

  • AI-generated reviews that pretend to be real customer testimonials are likely to create Fair Trading Act risk in New Zealand.
  • You can use AI to summarise genuine reviews, but the summary must fairly reflect the source material.
  • Keep records of the original feedback, consent, incentives and evidence behind any measurable claim.
  • Be careful with words like "best", "guaranteed", "everyone" and "number one" unless you can prove them.
  • Your website terms and review policy should explain how reviews are collected, moderated and used in marketing.
  • An advertising claims review can help you catch risky AI-generated wording before it goes live.

If you'd like help with your AI marketing claims, customer testimonial wording, website terms or privacy documents, you can reach us at 0800 002 184 or team@sprintlaw.co.nz for a free, no-obligations chat.

Alex Solo
Alex SoloCo-Founder

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.

Need legal help?

Get in touch with our team

Tell us what you need and we'll come back with a fixed-fee quote - no obligation, no surprises.

Need support?

Need help with your business legals?

Speak with Sprintlaw to get practical legal support and fixed-fee options tailored to your business.