If you’ve been to a marketing or market research conference recently, you’ll have noticed that most of the content is either expressly or indirectly about artificial intelligence and how it applies to our field. Same for the trade press – it’s heavily focused on AI.
As with pretty much every other profession, AI tools have significantly changed market research. From writing a first draft of a questionnaire to streamlining analysis, they can be extremely valuable. As a qualitative researcher, I’ve found them to be quite helpful in certain, specific situations. For instance, they’re great for automating repetitive tasks and are also very good (probably better than humans) at identifying latent themes in qualitative data.
As indispensable as these tools are, it’s important to remember what they can and can’t do. AI pioneers like Judea Pearl and Gary Marcus have pointed out that the term ‘AI’ is something of a misnomer, as artificial intelligence isn’t truly ‘intelligent’ in the most commonly understood senses of the word.
Rather, AI tools are high-dimensional probability calculators. They are good at yielding output that eerily resembles intelligence and are amazingly good at finding correlations and hidden patterns. As such, these tools can help us come up with new, unexpected hypotheses for further investigation. However, it’s essential not to skip that ‘further investigation’ part.
Using technology to automate tedious, recurring tasks or to sift through a huge, disorganized dataset is a smart practice. But, if you let a microchip think for you, not only are you asking for something it can’t actually do, but you are making yourself – as a carbon-based life form – superfluous.
Drawing new insights from data requires thinking. Managing a conversation in a spontaneous manner that leads to unexpected, insightful findings requires thinking. For now, at least, people will need to continue to do these jobs. So, if we want to stay relevant as humans, we need to lean into some uniquely human skills: critical thinking, inference, and nonlinear reasoning. Remember, probability and insight are not the same thing. True wisdom comes from things like intuition and lived experience.
When it comes to AI, a lot of the conversation seems to focus on efficiency. I like saving time and money as much as the next guy, but truly transformative qualitative research isn’t about efficiency, it’s about inference and insight in the face of uncertainty. For now, at least, that’s what the human mind is for.
Probability Isn’t Intelligence.