Ellen Phiddian
Cosmos science journalist
Do you ever laugh at an inappropriate moment?
A team of Japanese researchers has taught a robot when to laugh in social situations, which is a major step towards creating an android that will be “like a friend.”
“We think that one of the important functions of conversational AI is empathy,” says Dr Koji Inoue, an assistant professor at Kyoto University’s Graduate School of Informatics, and lead author on a paper describing the research, published in Frontiers in Robotics and AI.
“Conversation is, of course, multimodal, not just responding correctly. So we decided that one way a robot can empathize with users is to share their laughter, which you cannot do with a text-based chatbot.”
The researchers trained an AI with data from 80 speed dating dialogues, from a matchmaking marathon with Kyoto University students. (Imagine meeting a future partner at exercise designed to teach a robot to laugh…)
“Our biggest challenge in this work was identifying the actual cases of shared laughter, which isn’t easy, because as you know, most laughter is actually not shared at all,” says Inoue.
“We had to carefully categorise exactly which laughs we could use for our analysis and not just assume that any laugh can be responded to.”
Read more: Chess Robot breaks child’s finger: “This is of course bad”
They then added this system to a hyper-realistic android named Erica, and tested the robot on 132 volunteers.
Participants listened to one of three different types of dialogue with Erica: one where she was using the shared laughter system, one where she didn’t laugh at all, and one where she always laughed whenever she heard someone else do it.
They then gave the interaction scores for empathy, naturalness, similarity to humans, and understanding.
The researchers found that the shared-laughter system scored higher than either baseline.
While they’re pleased with this result, the researchers say that their system is still quite rudimentary: they need to categorise and examine lots of other types of laughter before Erica’s chuckling naturally.
“There are many other laughing functions and types which need to be considered, and this is not an easy task. We haven’t even attempted to model unshared laughs even though they are the most common,” says Inoue.
Plus, it doesn’t matter how realistic a robot’s laugh is if the rest of its conversation is unnatural.
“Robots should actually have a distinct character, and we think that they can show this through their conversational behaviours, such as laughing, eye gaze, gestures and speaking style,” says Inoue.
“We do not think this is an easy problem at all, and it may well take more than 10 to 20 years before we can finally have a casual chat with a robot like we would with a friend.”