5 myths about AI in consumer insights

aytm logo icon
Posted Sep 28, 2023
Janel Hagaman

There are so many myths surrounding AI in business, market research, and consumer insights. Is it a job-stealing threat? A magic wand? A passing fad? Let’s dive in and attempt to separate fact from fiction. In this article, we aim to clear up common misconceptions around AI—specifically within the context of consumer insights, where many myths persist about what AI is capable of today and how it should be responsibly deployed. So let's address some of these common misconceptions to develop an informed perspective on how to responsibly leverage AI to understand our consumer.

Let’s align on the meaning of AI

AI has moved from science fiction to specialized tools, yet hype still obscures its pragmatic potential.

Artificial intelligence (AI) is a broad term that has come to encapsulate everything from basic machine learning algorithms to advanced generative models like Midjourney, ChatGPT, and more. But the broader cultural view of AI as human-like robots, or as the looming "singularity" has colored perception of the technology as fundamentally different and separate from human intelligence. These sci-fi-fueled depictions of artificial intelligence dominate the media and spread through pop culture, so it's no wonder there are so many misconceptions swirling around AI and what it's really capable of. 

In reality, the implementation of AI is less about human-mimicking robots and more about machine learning algorithms laser-focused on specific tasks. Virtual assistants understanding voice commands? AI. Product recommendations based on your purchase history? You guessed it—that’s also AI. But it doesn't possess true general intelligence or consciousness—yet. The truth is that most current AI operates within narrow bounds, excelling at specific tasks like language processing, computer vision, or prediction. True artificial general intelligence simply doesn’t exist. Still, AI has become deeply embedded in our everyday lives, powering recommendation systems, language translation, facial recognition, and so much more.

In the business world, AI is increasingly being used to supplement and enhance human capabilities, rather than replicate them entirely. This is especially true in market research, where AI stands to change the way researchers leverage insights technology to unearth deeper understandings about consumers. However, many myths and hype still persist about how AI should and shouldn't be applied.

Myth 1: AI will replace researchers

AI is the future collaborator, not replacement, for skilled researchers.

The pop culture notion of AI as human-mimicking robots and the idea of a "singularity" fuels the myth that artificial intelligence will inevitably replace human jobs and capabilities. This fans a fear that AI will completely take over market research roles and eliminate the need for human strategic thinking and creativity. We beg to differ.

The current reality of most AI applications focuses on simply automating narrow, repetitive analytical tasks rather than exhibiting true general intelligence. Sure, a large language model (LLM) may be able to spit out hundreds of decent survey questions, but it still takes a skilled researcher to develop a short and impactful 5-minute survey. Likewise, AI might be able to assist in survey analysis, sorting through data to identify trends, but it lacks the reasoning and judgment to interpret findings and make strategic recommendations. 

So rather than replace researchers, it seems like (for now, at least) AI stands to augment human capabilities—not replace them. Put plainly, the idea isn’t to have AI do the thinking, but to execute on its specialized and narrow automation. By handing off tedious, error-prone tasks to AI, researchers can spend more time on high-level strategy and impactful analysis. 

“Many assumed that the promise of technology was to free up researchers from operational and administrative tasks in order to focus more on consultative and strategic functions. However, since 2018, we have tracked how insights professionals allocate their time to tasks, but seen no evidence of such a trend. Instead, it appears technology has increased bandwidth and throughput without reallocating tasks. Perhaps advances in AI may finally transform the daily life of researchers, but so far, we see no evidence of that shift happening.” - 2023 GRIT Insights Practice Report

Our key takeaway here is that it's important to recognize the current limitations of AI technology while embracing how it can boost researcher productivity and effectiveness. AI should be seen as a collaborator and force multiplier for people, not a replacement for researchers. 

Myth 2: Everyone is using AI

The reality of AI adoption rarely matches the hype.

There’s an assumption that artificial intelligence has already seen universal adoption—there’s so much hype! But the reality is that while AI usage is growing rapidly, it is far from ubiquitous at this point. According to a Mckinsey Study, “less than a third of respondents continue to say that their organizations have adopted AI in more than one business function, suggesting that AI use remains limited in scope.”

The relatively slow uptake among many organizations comes down to continued misunderstandings as well as practical challenges surrounding AI. So why the misunderstanding? Well, some companies believe AI is still largely unproven or unnecessary for their business goals. But the truth is that even basic machine learning has become accessible and impactful for most organizations.

When it comes to market research, we can see that some form of AI adoption has taken place several years ago. At aytm, we use artificial intelligence in a number of ways, including to help with image sorting by similarity, allowing researchers to save time combing through thousands of image responses. But the idea of semantically asking a computer to find correlations in data sets is very exciting, and something that needs to be executed correctly. 

This highlights some of the other barriers to broader adoption, including all legal and ethical concerns that come with generative models. Many of today’s models face accusations of being trained on copyrighted data, and many of the big LLMs are open about their data collection practices. These issues raise important questions around intellectual property and data privacy, and highlight the need for cautious deployment in market research contexts. 

So while AI adoption is accelerating, it has far from reached its full potential. By viewing AI realistically rather than through hype or skepticism, more companies can deploy it responsibly to enhance capabilities. Companies moving quickly today will have a competitive advantage over laggards who fail to realistically adopt AI, but they also need to take the risks seriously.

Myth 3: AI is a magic wand

AI is not a magical cure-all, it requires care and oversight

There’s a tendency to see any sufficiently advanced technology as indistinguishable from magic. Remember when Siri could captivate a crowd by telling a joke? Well, when it comes to AI, the illusion lies in believing it can magically solve your problems  without a concerted effort. Sure, AI provides powerful tools, but it’s not some golden goose.

Like any technology, AI can be applied in responsible or irresponsible ways. Unchecked, AI systems will absorb and amplify biases in data used for training models. But with thoughtful oversight and testing, AI can be crafted as a tool to increase research objectivity. The key is recognizing both the opportunities and limitations.

For example, a natural language processing (NLP) model totally help remove biased language from survey questions, but only if it is properly trained and validated on inclusive data samples. And while AI isn’t inherently biased, blindly applying can lead to problems. Achieving a fair and equitable AI requires embracing it as a set of versatile tools, not just a single spell.

That being said, organizations hoping for some quick magical fix to bias issues will be disappointed. But companies investing time to validate, monitor, and govern AI to align with ethics and inclusion best practices stand to gain a powerful advantage. It’s clear that an opportunity exists to wield AI tools responsibly, but we still need humans to train and nurture them. 

Myth 4: Generative AI is too new and risky to invest in

With the right governance, and agile approach, emerging AI enables smart risk-taking.

Gaining control of how AI is integrated into market research calls for cautious optimism. Organizations are curious, but also concerned. And at aytm, we’re very intent on ensuring our clients are protected—with features that will allow clients to discontinue any use of these new solutions by choosing to opt out at an account or a survey level. 

On the other side of the equation, suppliers would need to ensure that any experimentation of AI takes place on a private and highly-protected server, which  promises that our data can not be shared with or seen by external parties. This allows for  safe creation of exclusive data repositories for particular accounts that remain shielded from any external parties within a given framework. The results generated by AI solutions to be fully compartmentalized and precisely tailored to the needs, preferences, and prior history of a particular client. This means we would never allow the wording of a question for client A to be suggested as wording to client B, particularly in cases where such wording would be proprietary or typical for client A and not the industry at large. 

One way to look at implementation of generative AI is through iteration. For example, a research team could choose to deploy an AI image generator to develop a parody between sketches for concept testing. Rigorous testing at each phase could provide the necessary confidence to expand the use of this tech judiciously over time, and partnering with legal and security teams can help establish governance practices to ensure the experiments are conducted safely. So rather than let the newness of a technology delay adoption, it can be used as a reason to innovate iteratively and build capabilities over time.

Myth 5: The AI boom is already over

We’re still in the early stages of understanding how AI can transform the world.

Given the explosion of media coverage and content surrounding generative AI in recent years, it's tempting to think the hype cycle has already peaked. But in reality, we’re likely still in the early stages of generative AI's technology adoption curve. True integration into business practices is just beginning.

According to research by Mckinsey “40 percent of respondents say their organizations will increase their investment in AI overall because of advances in generative AI” This signals that we’re likely entering the "slope of enlightenment" period, where companies move beyond proof-of-concepts into practical AI deployment. But there’s still time for businesses to establish competitive advantage.

Rather than dismiss AI as a fad that’s already passed, a strategic outlook should seek to accelerate adoption responsibly in order to become a leader rather than a laggard. The bottom line is as the technology matures, we’ll see it transform from a novelty into a necessity. 

In the market research context, we can imagine using AI to expedite pre-flight checks, or serve as a copilot for survey authoring. And looking further out, possibly the most exciting modality of AI in market research would be to semantically ask for significant correlations in purchase intent between different consumer segments and receive statistically-backed responses supported by examples with your dataset. Of course, these responses would need to be validated by human eyes on verified data, but imagine the incredible advantages this would provide researchers. 

But avoiding the trough of disillusionment requires filtering hype and developing sustainable AI strategy. While the landscape will continue to evolve, organizations investing now in high-value applications of AI will be poised to reap benefits well into the future.

Moving forward

AI's future is bright, if we dispel myths with reality and temper excitement with responsibility.

When it comes to AI, sustainable results come from reality, not mythology. AI is not a magic wand that automatically solves problems like bias without diligent governance. But it also isn't to be feared as an uncontrollable force that will replace human judgment. 

The measured path is to carefully pilot applications that augment human capabilities and insights—to establish thoughtful oversight and validate against research ethics principles. We should embrace AI as a versatile set of tools that, with responsible design, can drive business value while advancing societal good.  

As we move forward, let’s continue to take this careful and agile approach to AI, debunking myths and setting pragmatic expectations as we move forward. Optimism is good. But let’s ground it in reality. There’s no doubt that AI will transform market research, but it’s up to us to understand how. For example: Did AI write this article? Nope—a human did. But we did ask AI for comment, and it happily provided us with the quotes at the top of each section. Could you tell?