Why Consumers Don’t Trust AI-Driven Recommendations

Data shows that there’s emerging skepticism surrounding AI in consumer decision-making. Why is there a disconnect and what should brands keep in mind?

Written by Thomas Peham
Published on Dec. 18, 2023
Why Consumers Don’t Trust AI-Driven Recommendations
Image: Shutterstock / Built In
Brand Studio Logo

We’re on the brink of a full-on revolution in the way consumers find and purchase products — at least that’s according to some AI advocates.

Through advancements in machine learning and AI, shopping will be rendered frictionless. Consumers will be presented with precisely the right products at the right times on the right channel. Guesswork and chance will be eliminated; conversion rates will skyrocket; and the complicated business of actually selling things will become a simple matter of fine-tuning your third-party AI.

It’s an appealing idea. However, as we often know, what sounds good in theory can end up very different in practice. We recently conducted a survey of 2,000 consumers and found that 85 percent of shoppers were not interested in using AI to help make purchasing decisions, and 60 percent said they would be unaffected by AI recommendations. On the face of it, consumers see little value in AI when buying online. 

3 Reasons Consumers Don’t Trust AI for Shopping

  1. Data privacy concerns: AI represents a more significant data privacy intrusion than social media collected data for marketing.
  2. Ineffective marketing channels: AI is being used as a solution to ineffective marketing channel strategies. A unified marketing strategy is more important than AI recommendations.
  3. Bad recommendations: A lack of data means most AI recommendations aren’t very accurate, leading users to doubt AI-generated suggestions.

Of course, we all know that AI, or at least machine learning, already plays a role in e-commerce, not least through back end automation and marketing personalization. However, these processes are subtle and don’t take the form most people now associate with AI, such as chatbots or image generators. So, why the disconnect? What is it about more overt uses of AI that seems to cause such antipathy? If AI is to truly revolutionize e-commerce, what can brands do to convince consumers that it is worth their while?


Why Don’t Consumers Trust AI for Shopping?

Looking back, the Cambridge Analytica scandal at Facebook (now Meta) represented a turning point in the relationship between consumers and technology companies. One that, per the Bipartisan Policy Center, “profoundly impacted the world of data privacy.” 

Post-scandal, consumers are much more attuned to how their personal data is being used and have only grown more concerned with each passing year.

Obviously, this has had a major impact on how consumers use social media: per a recent Pew Research Center survey, 77 percent of Americans have little to no trust in how social platforms might use their personal information. But it might come as a rude surprise to some AI boosters that the same skepticism applies to their field, too. According to that same Pew survey, 70 percent of respondents say “They have little to no trust in companies to make responsible decisions about how they use AI in their products,” and 81 percent believe their information will be used in ways they’re not comfortable with.

The numbers are somewhat lower in a recent Cisco report, but tell effectively the same story: 60 percent of respondents to its 2023 Data Privacy Benchmark Study expressed concern with the way AI is currently used by organizations.

Some of this might simply be the shock of transition, a brief fever that will pass with time. Famously, Facebook’s users panicked en masse when it first introduced its News Feed, feeling their privacy was being invaded. Fifteen years later, it’s their most important product and serves as the model for most existing social platforms.

But I suspect the transition won’t be so easy in this case. AI represents a much more significant privacy intrusion, and companies are rolling it out at a time when consumers are significantly tech-savvier in 2023. A recent FTC bulletin on consumer concerns with AI outlines some of the most persistent concerns, including biased data sets and the potential for getting scammed.

More on AIAI Bill of Rights: What You Should Know


Clunking Marketing Channels and a Lack of Data Mar AI

But the problems facing the mass adoption of AI in an e-commerce context go far deeper than security, to the very root of the industry’s problems today.

In Zoom meetings across the country, AI vendors are pitching their products as magic bullets: “Simply integrate our technology, and you’ll see your engagement soar.”

But here’s the thing. In e-commerce, as in most other aspects of life, quick and easy fixes tend to be illusory. They might make a significant difference, but only if every other element of your marketing strategy is in top shape.

Speaking from experience, I can say that many of the companies leaping to integrate shiny new AI tools would be wise to first take a look at their own websites. In fact, 43 percent of consumers feel a brand’s website is more important than its mobile app or social media accounts. 

Put otherwise: AI-facilitated recommendations will only get a brand so far if their various marketing channels are clunky, poorly integrated or otherwise not up to par. Consumers are recognizing that brands are prioritizing AI over the things they actually care about, like consistent, high-quality cross-channel content and interactions. 

Then we have the issue of data. What is often neglected in discussions of AI is the fact that AI can only ever be as good as the data brands feed into it. And right now, for too many brands, that data is disorganized and dispersed across multiple systems. 

According to a report from Hazelcast, 80 percent of companies globally are struggling to unify their data assets, drastically reducing the effectiveness of real-time recommendation triggers (not to mention deals and special offers). 

What this suggests is that consumer wariness about AI recommendations may be a consequence of the fact that AI-driven recommendations simply aren’t very good, because they’re operating off limited information.

More on AIWhat Is Relational Trust in Tech, and Why Do We Need to Build It?


Future of AI-Driven Purchases

Does all of this mean brands should entirely ignore AI? Of course not. It does mean, though, that brands should take actionable steps to modernize their CMS and to unify data from disparate channels. AI transparency is important here, too. Many of the concerns consumers have about AI-driven recommendations can be allayed through open, honest messaging around how their personal data is being used. 

Because while AI really might be the revolutionary technology its boosters are pitching it as, at the end of the day it will only benefit a brand if it is applied holistically — as a useful tool, not a “magic bullet” solution.

Hiring Now
Scythe Robotics
Artificial Intelligence • Computer Vision • Hardware • Machine Learning • Robotics • Sales