A regular part of Jon the Robot’s stand-up routine is to say he gets many auditions but few bookings because agents “always think I’m too robotic.”
If that line gets a laugh, he’s programmed to say, “Please tell the booking agents how funny that joke was.”
If it doesn’t, he says, “Sorry about that. I think I got caught in a loop. Please tell the booking agents that you like me … that you like me … that you like me … that you like me.”
Like any comedian, Jon reads his audience and responds to the feedback he gets. That’s what OSU robotics engineer Naomi Fitter programmed him to do. She took him to comedy clubs in the greater Los Angeles area, and then again to clubs in Oregon, to gather data on how humans and robots can use humor to interact more smoothly.
“Social robots and autonomous social agents are becoming more and more ingrained in our everyday lives,” Fitter, an Assistant Professor of Robotics at the OSU College of Engineering, told KTVZ-TV in an interview. “Lots of them tell jokes to engage users – most people understand that humor, especially nuanced humor, is essential to relationship building. But it’s challenging to develop entertaining jokes for robots that are funny beyond the novelty level.”
Human comedians develop their skills by performing at small comedy clubs before they try for the big time, Fitter reasoned, so she consulted with a pair of comedians on how to do field work. First, she booked 22 gigs in the greater Los Angeles area, which compared how Jon performed with a well-paced sense of timing to his jokes as opposed to simply rattling them off one after the other at the same pace. Not surprisingly, when Jon told his jokes the way an experienced human comedian does – pausing to allow the audience to react to one joke before moving on to the next – he got better results.
“In bad-timing mode, the robot always waited a full five seconds after each joke, regardless of audience response,” Fitter told the station. “In appropriate-timing mode, the robot used timing strategies to pause for laughter and continue when it subsided, just like an effective human comedian would. Overall, joke response ratings were higher when the jokes were delivered with appropriate timing.”
In a second round of ten stops at comedy venues in Oregon, Jon gave an “adaptive performance,” using not only timing based on how long the audience laughed at each joke, but adding “tags” to follow onto jokes which were well-received – what comedians sometimes call “milking” a joke. One finding from the second study was that a performance where Jon used “tags” wasn’t considered funnier overall, but individual jokes which Jon “tagged” were.
Besides helping to improve how robots are programmed for interacting with humans, Fitter hopes that her research might benefit human comedians by providing insight into how their performances worked – or didn’t.
It’s been said that dissecting a joke is like dissecting a frog. Perhaps Fitter’s techniques will yield a way of giving a joke an MRI.
By John M. Burt