To Relieve Holiday Stress, Techies Trot Out Artificial Intelligence

By now the journey to 2018 can feel more like a crawl than a mad dash.

There’s pressure to entertain family, reconnect with old friends and take that special someone on a memorable date. But a Richmond startup says artificial intelligence can solve the indecision over where to go and what to do. While some experts caution against placing exaggerated faith in artificial intelligence, early adopters are hoping for a more perfect holiday experience.

“Right now, planning fun isn’t very fun,” says Nate Marcus, the 35-year-old founder of a growing startup called OccasionGenius. “We want to make finding fun stupidly easy.” Local investors injected $500,000 in May alone and have committed for the long haul. “They’ve said, ‘This service will exist eventually, so why not us, why not now, and why not Richmond?’ They want to put Richmond on the startup map,” he says. 

Marcus says his mobile app creates what he calls personal genomes of users. First, users select from various categories or moods, such as active, offbeat or romantic. Boozy is Richmond’s most chosen mood, according to data. Next, users select from a slew of event types, such as open mics, restaurant pop-ups and festivals. To get a better understanding of each person, OccasionGenius pulls from Facebook to check events users attended in the past, and monitors browsing behavior on the site. According to one estimate, the company has logged 73,000 local users a year since launching in 2016.

Funny things happen when people look to have their desires affirmed, says Milos Manic, a Virginia Commonwealth University professor who specializes in machine-human interactions. If artificial intelligence is promoted as an authority, people tend to place more trust in its recommendations, even if the same suggestions are made by a friendly stranger.

“There are many definitions, and this field is changing continually,” he says of artificial intelligence. “But one common denominator is the ability [for machines] to learn.”

Not everyone gets warm fuzzies at the thought of machine learning. As artificial intelligence arrives in boutique and unthreatening forms, authorities such as Stephen Hawking and Elon Musk say that it could quickly learn to be nasty.

“Some fear is legitimate, like autonomous weapons, some fear is undue but should not be hastily dismissed,” says Matt Der, chief technology officer at Notch, another local artificial intelligence startup. “This topic warrants productive conversation. Certainly, there are ethical concerns in AI.”

A viral example came as recently as last month, when users recorded a digital assistant called Alexa making contentious statements about religious figures. Within a day, Amazon had altered Alexa to give Wikipedia responses, a move which observers found equally concerning.

“AI actually has the advantage that its recommendations are more objective than a human’s judgment, provided that no human biases were embedded into the model’s construction,” Der says. “But to the point, oftentimes a model’s suggestion should not replace, but instead inform, guide, or complement, a human decision.”

So should you worry about machines plotting to set you up on an awkward New Year’s Eve date? That’s not included here, Marcus says.

OccasionGenius’ secret sauce is NLP, or natural language processing, which sorts through events based on the description. Marcus says the software gets smarter as it goes, taking the frequency of words into account for an ever more accurate understanding of events pulled into the system. Events are then tagged, and more than 100 tags have accumulated so far. One of the most popular tags right now is, of course, winter holidays.

“We really geeked out on this,” Marcus says. “For example, is the event boozy like a wine tasting or does the event just have an open bar in the background? This helps us understand each event to a very specific degree, which no one else is doing.”

Marcus adds that his company, which has five full-time workers, won’t broadcast someone’s whereabouts on social media.

“We do not post anything to Facebook,” he says. “We think that’s weird. Our core values are transparency and empathy, so we also do not sell users’ information.” 

Just remember, if you’re struggling to plan a not-so-silent night, don’t give up on your gut, says Tom Arodz, another VCU professor who studies machine learning.

“AI may learn to never recommend a symphony to heavy-metal lovers,” Arodz says. “But just like with human instinct, it is often difficult to say why any particular recommendation is made.” S


WHAT YOU WANT TO KNOW — straight to your inbox

* indicates required
Our mailing lists: