One day soon, a menu may judge you.
You’ll walk up to a kiosk at a quick-service restaurant and a small camera will scan your features, recording your height, age, gender, and mood. Instantly, it will adjust its display, selecting meal options chosen just for you.
Once you have ordered and moved on, the person behind you will enter the menu gaze and the process will begin again.
That’s the idea behind new software from Raydiant, a San Francisco-based software company that plans to roll out its AI-powered kiosks by the end of this year. The intention, according to CEO Bobby Marhamat, is to create a personalized experience for customers, while helping restaurants increase sales.
“Our business started as a digital signage business. Customers [businesses] wanted to use analytics to create better experiences within their properties,” says Marhamat. Raydiant started developing AI technology in 2013.
Here’s an overview of how Raydiant’s kiosk works:
Raydiant’s technology, while not yet implemented with AI experiments, is used by more than 4,500 organizations around the world, from stores to gyms to casinos.
The companies themselves determine the criteria for age, gender and other characteristics. “Generally, the big brands make their own definition. Small brands follow big brands,” says Marhamat.
Raydiant has rules about the types of categories the AI can sort. Race and body size are prohibited. Gender, says Marhamat, is low on the list of sorting options brands prefer to use. “Most brands aren’t looking at gender right now. They’re looking at age group, time of day…different inputs. Age groups are categorized by facial features like wrinkles. L mood is subjective and based on data such as weather.” What has been identified is that if it’s raining and someone doesn’t like the rain, they walk disgruntled, the mood on their face is seen as discontent.”
Why Raydiant Says Tech Isn’t Scary
When Bon Appetit wrote about the new software, the magazine called it “scary” for obvious reasons. AI systems can be biased, playing on and reinforcing age and gender stereotypes. As technology advances, how personal data is used, whether anonymous or not, is a concern for groups advocating data privacy.
Marhamat, unsurprisingly, disputes these characterizations. Customers don’t have to worry about data privacy and face analytics because his company doesn’t log individualized data or sell it to third parties, he says. Users can also choose not to save their features by not using AI-enabled kiosks.
But the technology raises serious ethical questions. Food advertisements are powerful drivers of consumer behavior. Does Raydiant have a certain responsibility to guide its restaurant customers towards ethical marketing decisions? Wouldn’t it be immoral, for example, to effectively hide menu images of salads or fruit from young customers, the very people who are likely to be victims of fast food advertising?
Each brand that adopts Raydiant kiosks implements its own system, he says. “We don’t control it, the brands control it. It’s more like we give them the ability. Marhamat says Raydiant provides companies with privacy documentation to help them meet legal guidelines. On the backend, Raydiant says the technology is designed with privacy by default, meaning only anonymous data is processed and collected. Businesses can also enable blurring functionality in addition to analysis.
“This is just the start where people feel comfortable using this technology to improve their experiences,” says Marhamat.
Raydiant’s technology is designed for quick-service restaurants and is unlikely to take hold in the full-service restaurant space, Marhamat says. Still, Cameron Fraser, restaurant industry veteran and service and beverage manager at Flame and Smith in Prince Edward County, Ont., says he finds the premise depressing no matter where the product is. deployed. “It’s more the tyranny of the algorithm,” he says.
Algorithmic marketing traverses physical space
Raydiant’s tool is also part of a larger trend in retail, Marhamat says. Brands will soon routinely use facial recognition software to rank customers and cross-promote products that might fit their niche. “If you go to the yoga pants section, is it possible to advertise a yogurt that is in another section of the store?” he says.
Categorizing customers by demographic or type means that majority preference determines the advertising you get, even if you don’t fit the norm. Whether or not we are ready for the implications of these changes, online silos may soon be happening offline.
Fraser likens the smart menu kiosk to digital music services that stream hit singles, making an artist’s entire body of work or albums less visible. Your approach to food — like music, books, or movies — should be about exploration and expansion, he says, not based on algorithmic recommendations.
“The joy is in discovery and novelty,” says Fraser. “If not, what are you going to do?” Eat the same ham and cheese sandwich every day? »