Artificial Intelligence Will Soon Shop For You — But Is That A Good Thing?

We’ve all had bad department store shopping experiences. The aggressively cheerful salesperson. The unforgiving glare of the dressing room. The overstuffed racks of garments where none of the sizes fit, and the ones that do, don’t come in your favorite color.

The advent of online shopping has helped consumers gain more control over their shopping experiences. But digital purchases are often a gamble, too. You scroll through endless webpages to find the perfect boots only to discover your size is on back order for two months. And the items you purchase frequently disappoint: The jacket that looked so elegant on the website’s model looks awkward on your frame.

Retail prognosticators claim that artificial intelligence and other new technologies will offer shoppers salvation. In the not-so-distant future, armies of robots using retina recognition software (à la "Minority Report") will tailor their sales pitches to your preferences and price point. Voice-activated assistants and digital mannequins will help you to find just the right fit. Shopping from home will be a breeze too: Virtual reality headsets will allow you to “try on” clothes and sample items ranging from a tube of lipstick to a tennis racket. Two-day shipping? How antiquated. In the future, your package will arrive via drones in less than two hours. It may sound like science fiction but, in fact, many stores are testing these innovations and have plans to roll them out to customers.

Before you get too excited about this version of space-age shopping, you may want to consider whether these technological advancements are beneficial. It’s clear what retailers and brands gain: your personal information — including your measurements, spending patterns, individual habits and contact details. Indeed, retailers know more about you than ever before. And new methods of machine learning that allow companies to process large amounts of personal sales data, combined with beacon technology and insights from behavioral science and predictive analytics, help retailers target their wares with pinpoint precision.

This alone is worrying. We already rely too much on technology to make decisions for us. We presume falsely that these choices are good for us, or at least benign. But this is often not the case. Think about how often you outsource decisions and tasks to technology based on the assumption that it will improve your performance. When was the last time you mapped out an efficient route to an unfamiliar address without Waze or Google? Do you know the birthday of your closest colleague, or do you rely on Facebook to remind you? When was the last time you searched for cool new music without Spotify or Pandora?

On one hand, it makes sense to delegate these tasks to machines. Research shows that the conscious mind is “boundedly rational” and cannot equally process all of the stimuli in our environments. Relying on shortcuts frees our minds to focus on more important functions.

On the other hand, this cognitive cyborg reality can be dangerous. By unconsciously outsourcing more of our minds to technology, we become overconfident in the technology’s ability to be rational, objective and bias-free. The problem is those digital algorithms and applications were programmed by people who have cognitive biases and, not only that, have an objective to try to sell us stuff or make us think a certain way. Machine learning, the new frontier of data science, learns from human behavior, both good and bad. And as a result, we may be more susceptible to nudges and prods that might not be in our best interests.

By unconsciously outsourcing more of our minds to technology, we become overconfident in the technology’s ability to be rational, objective and bias-free.

Our outsourced minds need to adapt to this new world by consciously combating biases in technology.

Behavioral science has shown us time and time again how important it is to check our gut responses. This is no less important when these responses are mediated by technology. This new reality is motivating research that I’m undertaking at MIT; we seek to understand how sharing cognitive tasks with artificially intelligent machines can impact the economy, our health, our finances and society in general. The goal is to understand how humans can harmoniously incorporate technology while continuing to combat biases. Given human adaptability and our march toward “homo technicus,” we’re confident that this is achievable.

In the meantime, those harshly-lit department store dressing rooms and pushy salespeople don’t seem so bad, after all. Shopping anonymously has its charms.