We wrote last week about five education-technologies believed by industry experts to be potential game changers. The most radical, and the most distant, is affective computing. Here we’ll look a little deeper at the concept’s origins, developments, and future potential to education.
Affective Computing:
‘…is computing that relates to, arises from, or deliberately influences emotion or other affective phenomena.’
History
Like any advancement in technology or education, it’s important to note the influence from both philosophical and technical work, the former of which usually guides the development of the latter, which usually gets the most attention. Psychologist James Williams touched upon the subject of machines emotionally catering for and responding to humans in his 1884 work “What is Emotion”. Here, Williams proposed that there was a clear distinction between the perception of an event, and the response to an event in regards to what emotions were at play. Essentially James argued that emotions were simply bodily responses (the mind included in the body), and noted that intelligent machines would be fundamentally different for this reason. This work has since been cited and expanded upon across psychology, philosophy, pedagogy, and technology literature. The phrase “Affective Computing” was popularised by Rosalind Picard in 1995, who described the benefits of affective computing as not only being able to better assist human life, but to make more intelligent and creative decisions, on the grounds that decision making without emotions is just as flawed as decision making without it. Historic developments in AI such as IBM Deep Blue’s 1997 defeat of chess-champion Garry Kasparov have made normal the idea that on some level machines may already be more ‘clever’ than us. Intelligence is a multi-faceted concept, however, and there are still many areas where machines can’t match human thought. In terms of raw computational or what you might think of as mathematical power, there would be few remaining doubters of computer’s superiority. What hasn’t become so mainstream yet is imagining the possibility of computer’s being more emotionally receptive than us.
Developments
Google DeepMind’s defeat of world Go champion Lee Sedol stunned the tech-world recently. This was because the board game of Go (payed for millennia and once considered one of the four requirements for a true ‘gentleman’ in China) was considered impossible for machine’s to play, because it was considered part-science, part-art. The game is said to require deep human intuition and comprehension of asymmetrical, imperfect artistic forms – something a computer can’t traditionally be programmed to understand. What set DeepMind apart from other attempts at AI was that the development focus was on the machine learning from deep neural networks by means of watching what humans do. Interestingly, the programmers themselves were baffled by some of the moves played by DeepMind during the matches and could only suggest that the machine had learnt itself, in its own way. Looking at things from the perspective, perhaps very soon machines will be on par with human at an emotional level, albeit without programmers and the wider world being quite sure of how it is they are.
One project from MIT Media Lab’s Affective Computing group is taking a more traditional route. Researchers are exploring the use of wearable biophysical sensors to collect longitudinal data on how the environment and human experiences effect psychology. The main thinking here is that the human memory is not fine-grain enough to collect, interpret and pattern-match factors that induce different emotions, accounting for variables. Perhaps by analysing the shared emotional experiences of many humans, some powerful insights will emerge regarding emotion-cues.
Future Potential
The field of affective computing is generally split into two areas. The first is in detecting and recognition emotional information in humans. Here, through the use of powerful data capture and interpretation techniques, machines seek to gain a deep, quantified understanding of human emotions. An example of this would be in an eLearning context, a machine adjusting or altering its presentation of information, or style of presentation in response to perceived emotions from the learner – perhaps frustration, boredom, or interested. The second field is the design of machines proposed to exhibit either innate emotional capabilities or some variant of convincingly similar emotions. While most of the technology is already in place to research and explore the first area, lately attention has been turning to the second, which is likely to have broader implications for the future of society. Coupled with best-practices in psychology and therapy, emotionally capable computers may bring expensive mental health benefits to the mainstream, which itself would have dynamic effects across broad swathes of society. Particular ideas in play currently involve the use of computerised pets and the well-documented mental-health effects a live pet has on its human owners, as well as the use of an emotionally sympathetic, albeit private and safe computer interacting with an autistic child, who may feel more comfortable in the presence of a computer.
The two fields are likely to continue developing in tandem, and the effects this has on the philosophy and technical delivery of education will be astounding.