POWER > Trends > IBM Research

IBM Explores How Affective Computing Can Benefit Society


Q. How would it sense, over the phone, if I’m angry?
We’re currently looking at two things. We’re interested in voice—the tone, the pitch, the culture—to detect the emotion. But we’re also experimenting with other ways of detecting emotion. We’ve been working with an IBM partner doing facial expression analytics. The software analyzes facial expressions from camera input and in real time knows what those expressions mean.

Q. What are the potential applications for this type of technology?
There are a lot of applications. You can look at how people interact with the content of websites and applications and measure their emotions to improve the user experience. It could also be used to create a system that measures the emotions of people who are driving a car and have the behavior of the car adjusted depending on whether the driver is angry or distracted. The same type of technology could be used as a safety application for airplane pilots.

Q. Will computers become better at reading emotions than human beings?
There are two major types of emotion—primary emotions, which are mostly physiological and easier to detect, and secondary emotions, which are related to our interactions as social beings. I believe we will become more accurate in capturing these primary emotions. Fear is a primary emotion. It’s an experience that’s usually not controlled. On the other hand, disappointment is a secondary emotion. It’s much more complex. Disappointment involves goals and discovering that a new event may create problems as you attempt to reach them. To identify that you’re disappointed, a computer program needs to understand what your goals are and what new event created a disruption in achieving those goals.

Think about an emotion like shame. It’s a typical human emotion. But for a computer to understand that somebody is experiencing shame, it needs to understand the social and environmental context. It’s very cultural in some ways. When you ask whether computers are going to be better than humans at detecting emotion, it’s really dependent on the individual emotional situations.

Q. Do you think affective computing will become, or already is, a component of artificial intelligence (AI)?
I am convinced it is. We’ve developed a chess-playing computer, as you know. People have been working on that for a very long time, and IBM is part of that history. But a computer that plays chess doesn’t take into account the emotions of the human opponent. I’m not a chess player, but I think when humans play chess, emotions must have a role in how they play. Maybe a future chess-playing computer will understand the opponent’s emotions and use that against him. That’s probably the most obvious application I can think of, but I think this type of technology definitely has a big role in the future of AI.

Q. As a philosophical aside, do we really want machines evaluating our mental state?
There are a lot of ethical issues involved, and there’s a whole group of people looking into these. On one hand, you could say, “Well, it’s no different than a computer analyzing what I buy and then sending me further promotions, like any recommendation system.”

However, people tend to think of emotions as being more private, more related to their internal state. Sometimes people try to conceal their emotions and wouldn’t like a computer trying to detect what they are feeling. That’s why we’re looking at applications where the detection of emotions can help people, give them support by measuring their emotional state or give them better service or advice, like in an educational setting.

Jim Utsler, IBM Systems Magazine senior writer, has been covering the technology field for more than a decade. Jim can be reached at jjutsler@provide.net.

comments powered by Disqus



2018 Solutions Edition

A Comprehensive Online Buyer's Guide to Solutions, Services and Education.

IBM Systems Magazine Subscribe Box Read Now Link Subscribe Now Link iPad App Google Play Store