Autistics have more difficulty picking up on social cues like knowing when a conversation lasts too long and a device is being developed to help according to this article by NewScientist.com. The device is called the “emotional social intelligence prosthetic” (ESIP) and is being developed by MIT students El Kaliouby, Rosalind Picard and Alea Teeters. In short ESIP is a small camera with special image recognition software. The person with autism would put this on and it would analyze the reactions of the person being talked to and begin to vibrate if the machine detected boredom or irritation.

The curious thing to me about the device is that during the study phase it proved more difficult to detect the emotions of ordinary people than actors.

The software picks out movements of the eyebrows, lips and nose, and tracks head movements such as tilting, nodding and shaking, which it then associates with the emotion the actor was showing. When presented with fresh video clips, the software gets people’s emotions right 90 per cent of the time when the clips are of actors, and 64 per cent of the time on footage of ordinary people.

Shouldn’t it be the other way around? I mean, shouldn’t professional actors be able to fool machines with their emotions better than ordinary people? Or does their overacting make it easier for machines to figure out their emotions?

Let’s put aside machines and autism for a bit and talk about difficult social cues. Boredom is more challenging to interpret than irritation. I have trouble understanding if my wife is bored sometimes and we’ve been married almost 20 years but have absolutely no problem telling when she is irritated. There is this icy type of stare that could cut diamonds in half. Happiness, sadness, anger, frustration, all fairly easy emotions to interpret. Boredom is a bit more tricky because some people feign interest out of courtesy. Maybe that’s what trips up ESIP with ordinary people.

When/if these MIT students ever fully develop the ESIP they should explore offfering as a supplement for podcast and videoblogging viewers. When the listener or viewer starts to get bored or irritate a negative vote could be sent back to the creator so they can better detect the weak spots in their production. Auto content feedback and voting. For that matter, keep it going and expand into text. How about an RSS reader that while we read automatically skips past boring posts and remembers which publications had the most boring content and automatically unsubscribed.

Someday.

• Clocky the hiding robot alarm clock
• Voice actors of the Simpsons on Conan [video]
• Apple claims 4.7 percent of US market share
• Canadian dollar worth over $1.03 USD
• Simpsons voice actors feel they are underpaid, episodes will be trimmed in upcoming season
• Complete freeware solutions