New Zealand still doesn't have an AI-specific law. On Tuesday 18 November 2026 the Commerce Commission released guidance setting out how the Fair Trading Act 1986 and the Commerce Act 1986 apply when a business uses AI. The Commission says it isn't regulating the technology. Its interest is in what AI does to consumers, competition and markets. For most owners and IT leaders, the practical message is that any tool producing customer-facing words, prices or decisions now has a regulator's checklist attached to it.
Deputy Chair Anne Callinan's main line was: "Whether a decision is made by a person, an algorithm, or an AI system, the business remains accountable for the outcome." The release also says boards should be able to see how AI is used across their organisation, and it asks senior leaders to treat AI as a governance matter.
What the guidance covers
The guidance sits on a hub page with three detailed sections. The Fair Trading Act section lists eight risk areas: misleading conduct in AI-generated advertising, web copy and chatbot answers; claims without reasonable grounds; fake endorsements made with synthetic images, video or audio; AI-generated reviews and testimonials; "ghost shops" that use AI content to pose as New Zealand retailers; personalised pricing and offers; manipulative, highly targeted marketing; and "AI washing", meaning overstated claims that a product is AI-powered. The Commission also says AI output doesn't count as evidence that a claim is true. Marketing teams that draft product claims with a generative tool should note that point.
The Commerce Act section covers cartel conduct carried out by algorithms, misuse of market power built on large-scale data, and acquisitions of AI start-ups made to remove a future competitor. Its examples include a cloud provider that requires customers to take its AI products as part of a wider package, and AI pricing tools that set prices low enough to shut out new entrants.
The algorithmic collusion section is the most operational of the three. It warns that shared pricing platforms, data intermediaries and AI services can carry competitors' confidential information into a business's recommendations. It also says buying the tool from an independent provider doesn't automatically protect the buyer from liability. The section raises self-learning systems too, noting that software with little human oversight could learn to match rivals' prices without anyone agreeing to anything.
Why the penalties make this a board item
The penalties are already in force, and they're large. A Fair Trading Act breach can cost a business up to $600,000 and an individual up to $200,000, and each breach can attract a separate penalty. Under the Commerce Act, a business faces the greatest of $10 million, three times the commercial gain, or 10% of turnover, and individuals involved in cartel conduct face up to seven years' imprisonment. A chatbot that gives a wrong refund answer to thousands of customers could become thousands of misleading representations.
The timing also matters. The 7 November election left the coalition one seat above a majority on preliminary results, and official results are due on 27 November. Any new AI policy will take months to form, but the Commission's guidance applies to conduct now. Callinan told Reseller News that regulators will keep considering whether they need further tools. Businesses shouldn't treat the absence of an AI Act as a grace period.
What it means for NZ businesses
Our reading is that the guidance turns AI governance from a voluntary framework exercise into evidence a business may have to produce for an investigation. If the Commission asks why a chatbot quoted a wrong warranty term, or why an automated pricing tool moved in step with a competitor's, the business has to show how the system was configured, who reviewed its output and which data fed it. The collusion section says plainly that good records help show decisions were made independently.
Retailers, e-commerce operators and service businesses running customer chatbots carry the most immediate Fair Trading Act exposure. Pricing exposure is wider than many firms assume. Distributors, accommodation providers, fuel retailers and building suppliers often buy pricing or revenue management software from vendors that serve several competitors in the same market, and the guidance names that arrangement as a risk. Software companies and resellers that describe their products as AI-powered need to be able to back up those descriptions.
The guidance doesn't publish new thresholds, enforcement priorities or case examples, and the Commission hasn't said when it expects to take its first AI-specific action. The Bell Gully analysis from earlier this year had already reported that the Commission was monitoring AI-related competition risk, so 18 November is better read as a formal statement of a position the Commission was already taking.
Practical next steps
- Build an AI register that marketing and pricing can't skip. List every tool that writes customer-facing content, answers customers, sets or recommends prices, or screens applications. Include features switched on inside existing platforms such as CRM, e-commerce and ERP systems, because these rarely go through procurement.
- Test customer-facing AI before and after go-live. The Fair Trading Act section calls for regular testing of chatbots. Write test scripts around refunds, warranties, consumer guarantees and pricing, run them after every model or prompt change, and keep the results.
- Question your pricing vendors in writing. Ask what data the tool uses, whether other customers' data shapes your recommendations, how data from different clients is kept apart and what safeguards exist. These questions come directly from the Commission's vendor checklist.
- Keep a person accountable for automated pricing. The guidance expects staff who understand the system and review its outputs critically. Name an owner for each pricing tool and set the conditions under which that owner overrides it.
- Audit your AI claims. Check every "AI-powered" statement on your website, in proposals and in product sheets against what the product does, and file the evidence alongside each claim.
- Report AI to the board. A short quarterly report covering the register, test results, incidents and vendor answers gives directors the visibility the Commission expects.
- Attend or watch the webinar. The Commission and the AI Forum are running a free session at 12pm on 24 November.
Firms that already run an AI governance framework can add these items to existing controls in a few weeks. Firms that haven't started should begin with the register and the chatbot tests, because those two items cover the most likely complaints. Our AI consultancy team helps NZ organisations set up these controls. Where pricing or quoting runs through automated workflows, our process and automation work covers the logging and approval steps the guidance asks for.
Sources
- Commerce Commission issues AI guidance for New Zealand businesses (Commerce Commission, 18 Nov 2026)
- AI guidance for businesses (Commerce Commission)
- AI and the Fair Trading Act (Commerce Commission)
- AI and the Commerce Act (Commerce Commission)
- Algorithmic collusion and the Commerce Act (Commerce Commission)
- AI does not change competition and consumer protection laws, Commerce Commission warns (Reseller News)
- ComCom issues AI guidance for NZ businesses (FMCG Business)
- Algorithmic collusion and AI cartels: risks under New Zealand's Commerce Act (Bell Gully)
- 2026 New Zealand general election (Wikipedia)
Want to know what this means for your organisation? Our team in Auckland can show you your top three opportunities in a free 30 minute assessment.
Get my free assessment How we help