Human beings are social animals. As part of our neural wiring, humans assess each other’s emotional cues on a subconscious level without realizing it. However, some people, such as individuals with autism, have more trouble interpreting these cues. Some cues are difficult for everyone to read. Other cues are even intentionally misleading. Missing from most definitions of artificial intelligence (AI) is this subconscious emotional intelligence. If AI is ever to successfully work and communicate with humans on a peer basis, it needs to take human emotion into account. Technology that helps people recognize and read emotions also has the potential to help humans communicate more effectively with each other.
Emotion recognition technology identifies images of prototypical facial expressions of such universal human emotions as joy, surprise, sadness, disgust, fear, and anger. Emotient, an AI startup that was recently acquired by Apple, has developed software called Facet that reads emotions in pictures or video frames with resolutions as low as 40 by 40 pixels. When Facet analyzes video sequences, it is able to track emotions over time and even catch the “micro-expressions” that humans cannot control and are generally unaware they are signaling.
Since the uses of emotion recognition technology are seemingly limitless, Emotient is not the only player in this space. Emotion recognition software has been used by advertisers to assess consumers’ interest in products and their response to marketing techniques. It is being incorporated into video games to make them more emotionally intelligent and appealing. It can assist security personnel to identify individuals who should be monitored more carefully in sensitive situations and settings. Eventually, it may even serve as a different translation layer and help in diplomatic scenarios where cultural misunderstandings can lead to dangerous miscommunications.
The benefits extend to medical applications, as well. Emotion recognition software has been developed into a game that trains children with autism to interpret facial expressions. It is being cultivated to identify and treat depression. And it can be used by medical and mental health professionals to gauge discomfort, pain, and other feelings in patients who are not able to communicate verbally.
But the technology is not without controversy. Critics find some of the advertising uses manipulative. The technology can also be used for deception applications and covert applications that seek not to promote better communications but impair it. People who are concerned about privacy may also find the reading of micro-expressions to be invasive. Also, as with many deep learning and recognition algorithms, emotion recognition technology is vulnerable to scamming and hacking.
According to Tractica’s research, emotion recognition is just one of many promising new AI technologies that will emerge over the next 10 years. The diversity of AI applications will create significant opportunities in large and small market segments alike, and according to our Artificial Intelligence for Enterprise Applications report, spending on AI will increase to $11.1 billion by 2024.