Microsoft Develops A.I. That Understands Laughter Enough to Judge Comedy

The modern-day arbiter of quality at open mics is the humble applause-o-meter, which is a good way to get the room hyped but often less than accurate. Fortunately, future generations of hopeful stars will benefit from a more impartial form of assessment enabled by A.I. Indeed, Microsoft has already developed an artificial intelligence capable of detecting how hard a person is laughing that could easily imagine being used to judge a comedy show.

The company exhibited its face-scanning system in August for the Laugh Battle exhibit at the National Comedy Center in Jamestown, New York. In the exhibit, the A.I. was used to determine which comedian was best at delivering their jokes. Think of it like a much less-snarky Simon Cowell that tries to rate how hard you’re laughing.

This is #13 on Inverse’s list of the 20 Ways A.I. Became More Human in 2018.

Technology like this demonstrates how A.I. can be used to pick up on nuanced human expression. Microsoft’s corporate VP for A.I. marketing, Mitra Azizirad, also says its also a way to open up the benefits of machine learning to a wider audience.

microsoft laugh battle
A participant scores a point after making her partner chuckle.

“We are making A.I. accessible to everyone, expanding it beyond the world of developers and data scientists to every person – especially in ways that are universally understood and touch the heart,” she says in a statement. “Nothing does that better than laughter.”

Laugh Battle featured more than 100 pre-scripted jokes that players could use to try and make one another cackle. The system’s Face API uses a deep neural network trained on more than 100,000 labeled photos to judge people’s expressions on a scale that ranged from sadness and contempt to happiness and surprise. The A.I. then awards points to the participant that made their partner keel over the most.

This is proof that A.I. can be used to bring people together while also being used to augment human ability. But perhaps most important of all, it’s also an example of how A.I. is getting ever closer to mimicking even the most intrinsically human characteristics like knowing what’s funny.

Media via Microsoft