American start-up company, Affectiva successfully built an AI learning system that can recognize human emotions by analyzing facial expressions, is now building on their technological successes. With emotional recognition squared away, Affectiva’s new project aims to create a Human Perception AI, artificial intelligence that “understands all things human.” Although that may sound like a moon shot to some, investors are taking the endeavour very seriously: the company just raised $26M in their latest funding round, bringing their war chest to $53M.
Emotion AI is straight-forward enough: it leveraged existing studies of emotional “tells” that manifest as microexpressions – minute facial expressions that most people subconsciously read – and used phone or computer video cameras to capture these expressions as people were exposed to various stimuli, then interpreted those expressions as emotions. You can actually try the technology out on their Website: when you realize just how accurately the system can capture even the slightest facial expression, you may want to cover your camera with electrical tape.
Human Perception AI, however, is a much bigger goal. The name alone suggests creating a machine with the ability to perceive something as a human would, which leads us into the denser thorns of philosophical discourse. There is, however, a second possible meaning of the name: an AI whose entire purpose is to develop a perception of humans. A glance at the company’s site seems to suggest exactly that.
Building AI that understands all things human is no small feat. But with our proven approaches to deep learning, computer vision and speech science, and the massive amounts of real-world people data we continue to collect and annotate, we’re well on our way. What’s more, by taking a multi-modal approach to human perception — analyzing both facial and vocal expressions — were able to get a more comprehensive and accurate understanding of human states.”
Affectiva’s not playing games when it comes to developing this technology. They hold a staggering number of patents related to effective AI and the machine/software detection of human emotional states. The names and descriptions of these patents are fascinating – in some cases a little disturbing – and they suggest that Affectiva aims to collect a variety of physiological data that exceed “facial and vocal expressions” for reasons that go beyond mere “understanding of human states.” Among the patent names are”Mental state analysis using web services,” “Mental state well-being monitoring,” “Mental state analysis using heart rate collection based video imagery,” and “Pain analysis using electrodermal activity.”
Even if such technology didn’t serve to build upon the foundations of a surveillance state and were leveraged for the seemingly innocuous chore of marketing goods and services that you’d have a use for, would humanity be better or worse for it? It’s a question that should not only be asked by consumers, by anyone who aims to understand all things human.