Digital Snake Oil or Genuine AI? How to avoid AI-washing in Market Research

Generative AI has made its grand entrance. Its implications are dominating the headlines in tech circles and beyond. The research industry has wasted no time in harnessing its power, providing new ways to secure insight with agility. And early experiments have already shown that AI can help brands connect with consumers in a manner that is ‘naturally human’ at scale.

But as AI’s impact across market research expands, so is the hype – risking casting a shadow over its true value. Just as so-called greenwashing has led to consumer scepticism over companies’ ecofriendly claims, AI-washing could do the same for the market industry. So, when it comes to AI claims, it’s essential to channel our inner Sherlock Holmes and approach them with a discerning eye.

In this article, we roll up our sleeves and dive headfirst into the world of AI-washing, arming you with the know-how to separate the genuine AI marvels from the digital ‘snake oil’ lurking in the insights industry.

Be aware: ‘AI-powered’ everything

In my experience of exploring and experimenting with emerging technology at NOVA, MMR’s inhouse tech innovation team, I’ve come across multiple claims of AI in research tools and services that aren’t quite as “intelligent” as they’d have you believe. Through our collaborative adventures with several global CPG brands, we’ve been on a mission to pop the hood and see just how much of that AI magic is really under there.

You might be surprised at what we’ve found: tools or services listed in an “AI” directory, lack any trace of artificial intelligence within them, or marketing materials that toss around AI jargon like confetti without ever explaining how AI is actually used in the product. In one instance, conditional question branching (skip logic), was being sold as an ‘AI-powered’ survey.

‘‘AI-powered’ everything’ certainly comes with a price tag over ‘non-AI’ solutions. The result? A tangled web of misconceptions and false and overinflated expectations about what AI can deliver, making it more challenging for client brands to spot the truly innovative and useful AI ResTech gems hidden among the sea of AI-washed wannabes.

Investigate: Ask the right questions

This is not to say that AI needs to be everywhere, but when it comes to cutting through the marketing mumbo jumbo and getting to the heart of whether a product is genuinely AI-driven, it’s time to channel your inner detective. Asking appropriate questions and demanding some good old-fashioned proof will help you peel back the layers of sales-speak, revealing the true AI essence (or lack thereof) beneath. Start by asking them about the libraries and AI models they’re using. Do not settle for ‘proprietary tools’ or ‘secret source’ answers. You do not need a machine to predict the outcome of those claims.

Varied intelligence

Grasping the underlying principles of how AI tools function is very important, especially considering that not all ‘AI-powered’ tools are created equal—some indeed possess a higher degree of intelligence than others. For instance, in our comprehensive experiments with conversational AI, we’ve encountered chatbots of varying sophistication.

On one hand, we’ve seen chatbots that required programming to identify key topics or aspects of interest, enabling them to probe deeper into those areas with research participants using a predefined list. On the other hand, we’ve also come across more sophisticated chatbots capable of learning from conversations over time and engaging spontaneously with consumers. These advanced chatbots can hold ‘free-reign’ conversations while tapping into Large Language Models (LLM’s). While price tag can be the similar, the latter is a better investment option.

Experiment: Put the tech to the test

When delving into AI solutions, it’s essential to not only conduct small-scale pilot projects to assess the technology but also to collaborate with someone who genuinely shares your passion for technological exploration, rather than just trying to sell you the next big thing. Establishing a partnership with such a collaborator can help you navigate the risks associated with new technology adoption and provide valuable insights into the AI solution’s strengths and limitations, enabling informed decision-making about its practicality.

Comparing the AI product’s effectiveness against traditional research approaches allows you to evaluate its true value. When the AI technology outperforms conventional methods, you can trust its potential benefits. By fostering a genuine partnership, you can work together to uncover truly innovative and practical AI solutions.

AI Regulation

Emerging regulations might make it easier for all to discern whether a product or solution truly incorporates AI technology. We are seeing early signs that the rampant AI-washing is catching the attention of regulators in the US, and it’s likely that other countries will follow suit. The US Federal Trade Commission (FTC) published a blog post cautioning companies against making inflated claims about their AI capabilities. Its author highlights that not only do some AI-claimed products fail to deliver as advertised, but exaggerated claims about AI capabilities also fall into the realm of inaccuracy if unsupported by scientific evidence or only applicable to specific users or conditions.