Creepy New Tech at Wahlburgers Might Predict Your Order By Your Gender

A woman walks into a fast food restaurant. Then, a man walks into that fast food restaurant. They stand side-by-side at their own self-serve kiosk ready to tap in their orders, but they see completely different menus. The woman sees grilled chicken salads; the man sees burgers. No, it’s not a glitch. This is the latest effort by restaurants using tech to personalize the dining experience in the name of selling more food to more people—and it’s powered by artificial intelligence.

It sounds dystopian, but this technology is real, according to Restaurant Business, the industry publication that first reported it. Cameras on the kiosks analyze customers’ faces, allegedly categorizing them by sex, age, mood, and even how distracted they are when looking at the menu. Then the kiosks display menu items that seem more likely to appeal to them. The tech came up at this week’s National Restaurant Association show, the biggest industry event in the country, and, according to that industry publication, it’s already in usage at Wahlburgers’ restaurants, the burger chain started by celebrity brothers and heartthrobs of my youth Mark and Donnie Wahlberg.

Mark Wahlberg is an investor and advisor to Raydiant, the California-based company behind this tech. Per Raydiant’s website, it wants to make brick-and-mortar restaurant locations “more enthralling and more akin to the internet.” (Because the internet is such a wonderful place to hang out.) To help its cause, Raydiant acquired a company called Sightcorp in January that uses cameras to collect anonymous customer data—gender, age, attention level, and the amount of time a customer spends staring at the ordering screen. Put them together and, voila, restaurant menus that change based on what the customer looks like. Presumably, the company will offer this AI-assisted ordering to other restaurants, too. (Wahlburgers and Raydiant didn’t immediately respond to a request for more information.)

Enthralling and just like the internet, right? What could go wrong?

“The inherent and unconscious bias that exists in the world is amplified in AI, and therefore demographics should never be used in determining someone’s preferences,” one industry executive in the digital ordering space told me. (They’ve asked me not to share their name to avoid professional repercussions.)

It’s true: Artificial intelligence has caused some people real harm. In 2020, a man was wrongfully arrested because of a mistake in facial recognition software used by law enforcement in Michigan. Experts say that AI systems can exhibit bias when trained using data that underrepresents one gender or a certain ethnic group, which can lead to people having trouble getting a loan to buy a home, or even being treated correctly at the hospital.

Thankfully, choosing between a chicken salad and a burger isn’t so high stakes, and race is a factor notably absent from Raydiant’s list of demographic options. Still, sizing up a person’s sex—based on how they present to a computer—while they’re just trying to order a burger doesn’t sit right with me.

Of course, restaurants have relied on this kind of predictive tech for years. In 2019, McDonald’s spent hundreds of millions of dollars to acquire a technology company that promised to customize the chain’s drive-thru menu boards based on the time of day, popular items, or the weather. (Late last year it sold the technology to Mastercard for an undisclosed amount.) We’re also already fed plenty of personalized offerings through delivery apps and digital ordering platforms that suggest what we might like based on what we ordered or browsed in the past. But there’s a line to be drawn between promoting comfort food when it’s cold and rainy and “Looks like this bleary-eyed mom needs some coffee!”

That said, diners are conditioned to put up with it all in the name of convenience—and some might even welcome it. As that tech exec told me, “Honestly there is so much complexity on so many menus now that there is a need for, ‘Tell me what to order, machine!’” It’s reasonable to expect this technology is here to stay, and will continue to evolve in order to squeeze the most cash from every order.

So I have to ask, creepy kiosk machine: What might this female restaurant tech journalist in her late 30s order for a quick lunch while distractedly minding her kid?

Trick question! We’re going somewhere else.