Leading AI research group the AI Now Institute has called for new laws to restrict the use of AI to detect emotions, saying the field is built on "shaky foundations."
Artificial intelligence has been used for some impressive applications in the past few years, automating tasks that would previously have required a human. AI is fantastic at classifying things into categories based on a library of known data points, but some of those working with the technology don't understand its limitations.
The AI Now Institute has released a report on the use of emotion-detecting "affect recognition" AI tech, which processes photos or video of a person's face and attempts to classify their emotional state. The technology is already on sale publicly in products used in interviews, and is used to detect deception by criminal investigators and insurance companies around the world -- the only problem is that it hasn't been proven to work.
AI has had a very shaky history in the recruitment sector, with tech giant Amazon abandoning its own AI recruitment tool when it became apparent that the AI was discriminating against women. Companies using black-box AI sytems for important decisions that affect people's lives could find themselves highly liable to discrimination lawsuits, as the firm has no way to prove that the AI didn't discriminate based on protected classes such as race and disability.
In its annual report, the institute has called for software using emotion-detecting AI to be banned in applications that can harm people's lives or reduce their access to opportunities such as jobs. The sector is estimated to be worth over £15bn currently and is continuing to grow, but the institute warns that the technology is built on "shaky foundations."
Co-founder Professor Kate Crawford told the BBC that studies show "no substantial evidence that people have this consistent relationship between the emotion that you are feeling and the way that your face looks," and that despite this the tech is being used everywhere, "from how do you hire the perfect employee through to assessing patient pain, through to tracking which students seem to be paying attention in class."
Source: BBC News, AI Now Institute
