Artificial intelligence (AI) continues to be a hot topic in 2023. And yet “AI” remains an ambiguous term with no single definition that is often used to describe a multitude of different tools and techniques.
“But one thing is for sure”, according to the US Federal Trade Commission (FTC), “it’s a marketing term”.
Just because a term is a little ambiguous doesn’t mean that it can’t be used in advertising – or that it can be used liberally to describe any tech-related product without regard to what that product exactly can and can’t do. Advertising can’t mislead the public, and businesses can’t make false or unsubstantiated claims about their products. In New Zealand these rules are enshrined in the Fair Trading Act 1986 (FTA) and enforced by the Commerce Commission (and there are similar standards in Australian consumer law enforced by the Australian Competition and Consumer Commission).
The US has its own set of rules enforced by the FTC. The FTC has recently posted a blog outlining key questions marketers should ask before talking about AI in advertising:
Are you exaggerating what your AI product can do?
If you can’t substantiate any of your performance claims, or your claims are subject to major caveats that you’re not disclosing, you may be misleading or deceiving the public (or be likely to).
Are you promising that your AI product does something better than a non-AI product?
Comparative advertising is fine provided the comparisons are clear, reasonable and accurate. And you need adequate proof to back up any claim that your AI product is better because of the “AI” aspect.
Are you aware of the risks?
You remain responsible for your product and any impact it has once it is put on the market. You won’t be able to escape liability for something that goes wrong by blaming a third-party developer, or saying simply that you didn’t understand the technology or couldn’t test it. Trying to do so by including one-sided exclusions of liability may fall foul of the unfair contract terms provisions in the FTA.
Does the product actually use AI at all?
If you claim your product is “AI-enabled” or “AI-powered”, be ready to prove it – the expectation that you create with such claims should be matched by what's “under the hood”. For example, can you really say that your product is AI-powered when all you’ve done is use an AI-based tool once in the development process?
These are also useful questions to ask in the context of New Zealand law, and are likely to be factors that the Commerce Commission and the Courts will consider in assessing any FTA claim relating to AI advertising.
Last year, we discussed how the Court assessed Bunning’s “lowest prices” claims and taglines in the context of the FTA. You can read more here. The key takeaways from that case will also apply to any AI related claim: the Court will consider the specific facts that relate to the AI product and how and where the AI claims are made (including any language such as “AI-powered”), and will also look at evidence of consumer perception and confusion.
AI is here. But how the technology works – and the limits of what it can do – are not yet well understood by consumers. It’s important for advertisers not to obfuscate or over-promise on what their AI products can deliver. Businesses will need to be aware of the risks and take a reasonable approach when referring to AI in their advertising to avoid violating the FTA, and should have good systems in place to back up any representations they make about AI powering their products.
