Words are only the tip of the iceberg when it comes to communication. The tone of your voice and your fleeting facial expressions are essential tools when expressing your emotional state and feelings. Current technology eliminates these elements, leaving only words and a collection of emoji to get your message across. And yet, what technology takes away from us, it sometimes willingly brings back in the form of artificial intelligence. Silicon Valley has found a solution to the problem it has created: emotion-recognition technology. With this, your phone or other devices can interpret your mood based on your facial expression or your physiological data. Say hello to Siri, your new therapist.
Story by: Olivia Miossec
Technology: Machine-learning its way into your heart
Imagine coming back to your home. You open the door and multiple wireless signals shoot through you. They measure your heart and breathing rate. These data are fed into a ‘router’ that interprets your emotional state: excited, happy, angry or sad. This ‘router’ is the current ‘baby’ of researchers at the MIT’s Computer Science and Artificial Intelligence lab. It is known as Emotional Quotient (EQ)-Radio. This technology is made possible with machine learning. Through the collection of physiological and emotional data sets and complex learning algorithms the software can then be trained to predict a person’s emotional state based on body function data only.
This innovative technology, however, is limited only to four emotions. To go beyond this, technology must be able to read our biggest emotion interface: our faces. This is the realistic ambition of Affectiva, another company born out of MIT research. Their ‘emotion-sensing and analytics’ technology can read the subtle expressions on your face and translate these into accurate emotional insights. The software uses ‘deep learning’, a more complex form of machine learning. It mimics the layers of our cortical neurons. It breaks down faces into pixels, analysing the facial expression in each, weighing the pixels against one another, and finally merging them into a single emotion. It can recognize subtle emotions we would not even notice. In our defence, we were not exposed to 3.9 million faces from over 75 countries as part of our ‘training phase’.
In fact, our situation is quite the opposite. One study found that children who had spent five days with their smartphones were much worse at identifying various facial expressions than children who did not have access to their devices. So whilst technology is becoming more ‘emotionally intelligent’ every day, we are willingly letting ourselves be robbed of our innate abilities.
Emotion recognition: A double-edged sword?
What is the point of such technology? In the case of the EQ-Radio, its creators believe it could be of great value in the healthcare setting. It could monitor a patient’s mood alongside other physiological parameters and thereby detect feelings of depression and anxiety, known to impede a person’s recovery. As for Affectiva, they have already vowed to help the ‘emotionally blind’ with their technology. Emotionally-intelligent devices could be a breakthrough for individuals on the autism spectrum who have difficulties with ‘reading’ other people. This is the current work of the organization ‘Autism glass project’ that uses Google glasses paired with emotion recognition technology. The glasses record and analyse faces in real time, providing the wearer with information about people’s emotions, thereby improving his or her read on the social environment. Whilst the study remains in its infancy, the results so far have been promising.
Of course, these devices also have much less noble applications. At home, the ‘router’ could analyse your mood in response to certain commercials, movies or songs – and sell this data to companies. Similarly, if your phone can read your facial expression, it may transmit it to your friends and family to enrich connections, but it may also transmit it to corporations. Indeed, Affectiva already works in close collaboration with CBS and Kellogg’s.
Your Google searches and WhatsApp conversations may be valuable assets to advertisers. But your emotions? That’s the holy grail.