Can computers detect emotion and predict how people will react?

The influx of AI into our everyday lives is bringing a raft of technologies that can analyze our emotions and predict how we will react.
Written by Eileen Brown, Contributor

No longer the realm of science fiction, computers can now recognize how we feel and analyze our real emotions - regardless of what we say.

Seattle-based AI start-up SilverLogic Labs has developed an emotion recognition technology that analyzes and detects human emotions in ways that can predict how people will react or behave.

The company claims that this technology can predict emotions and human motivators "more accurately than any other technology available."

Its algorithm looks at tiny, almost undetectable human movements and works out how that person feels. It can provide an understanding of "what emotions mean" in a given context.

(Image: Silver Logic)

The technology works by analyzing facial movements associated with a range of human emotions displayed in private settings such as interviews, focus groups, and public appearances like press conferences, or legal testimony.

It watches humans, second by second, and then it compares each time-slice of emotion with others to determine how people actually feel when watching the TV.

In the UK, the Gogglebox TV program shows a selection of viewers watching TV shows. We can see their reactions to each program and easily determine how they feel. Imagine how powerful this information could be to marketers who could watch us watching live TV.

The emotion recognition technology focuses on subtle 'micro-expressions' behavioral characteristics associated with joy, trust, fear, surprise, sadness, disgust, and anger and eliminates human bias associated with these characteristics.

It then compares each facial gesture to its data warehouse of images to create a report, which includes an 'emotional score.' This score can be used for market research or to measure the human sentiment behind a public appearance for journalism or law enforcement purposes.

It uses deterministic modeling to provide customers with accurate predictions of market performance

SilverLogic Labs says that this type of technology will add an extra dimension of security to intense security or defense environments.

Someone running toward you may have completely different motivations for doing so. They might be malicious, are approaching to warn you of danger, or might simply be asking for your help.

Read also: DBPower 4K camera hands-on: An affordable waterproof HD action cam | Meizu M6 hands-on: Good for gamers, consumers, and business users | Doogee BL5000 smartphone hands-on: A low-cost workhorse with curves

Border protection, airports and other transportation checkpoints, high-value interrogations, and other national security applications could all benefit from emotion recognition technology to accurately and quickly work out people's motivations.

Healthcare providers can analyze and detect if a person is suffering from diseases or afflictions such as dementia, or it can detect if a person has drug or alcohol addiction.

It could also analyze whether a person on medication for ailments such as high blood pressure has actually taken their medication.

It could also benefit the world of education. The technology could analyze student reactions to what the teacher is saying in order to make sure that they understand the lesson. This could reduce total learning time and improve efficiency in the classroom.

In the media industry, it could advance market research using focus groups watching TV pilots and watching their emotions.

Trying to capture and analyze emotional responses to the channels that a person can engage with can yield unwieldy datasets that are hard to analyze. Accurately predicting customer behavior is difficult.

This technology could help us in so many ways. Emotion recognition technology goes beyond beyond conventional biometrics such as our fingerprints. It can work out our emotions and detect how we will behave when we think we are alone, which sounds a little like a Big Brother world.

But marketers will capitalize on this information to deliver ever more targeted messages to us -- before we even realize we have indicated our desire.

Previous and related coverage

Most US workers want to see more AI and robots in the office

Although workers want robots in the workplace to give them more time to do their primary job duties, almost all still want the human touch in the office.

AI can now predict and guarantee the results of influencer marketing programs

Is influencer marketing just another PR tactic or can it really be scaled into a credible media channel with AI?

Sighthound Cloud brings accuracy to emotion and gender detection

Sighthound Cloud can recognize faces and detect emotion in your images or video streams, with more accuracy than Microsoft Cognitive Services.

Editorial standards