In twenty years, facial expression monitors in frozen food section will be able gauge your reaction to pizza rolls and maybe, possibly, offer a deal based on your smile (two for one?). The same technology could be applied to dating apps like Tinder or suggesting movies on Netflix. While it sounds a little creepy, emotion recognition technology might just be the key to making our digital world more human.
Enter companies like Affectiva, which are developing so-called “Emotion A.I.,” which the company defines as the intersection of human emotion and artificial intelligence. Affectiva has been awarded seven patents so far and has filed applications for 30. Its emotion recognition software is used in numerous sectors including advertising, robotics, and healthcare.
Combining psychological research on emotions with big data and machine learning, Affectiva representatives say the software has analyzed more than nine million faces from 75 countries to create the most comprehensive emotional A.I. on the market.
On Thursday, at Techweek New York, the company demonstrated its AffdexMe app that uses your phone’s camera to identify facial expressions. It’s kind of neat but doesn’t do much. It is a peek into the kind of technology Affectiva is opening up to developers, though:
Using 40 million emotional data points, the app analyzes images input from your smartphone to tell you just how much you’re really hating that new CBS sitcom:
As virtual reality and robotics develop, emotional A.I. will be critical to keeping the nuances of human experience a part of its designs. For example, Affectiva is already being used in the game Nevermind, which adjusts play based on the user’s level of fear.
Gabi Zijderveld, chief marketing officer for the Affectiva, tells Inverse that Affectiva is eyeing the burgeoning eSports industry and that she also believes the emotion-recognition software could improve dating and recommendation-based apps.
The company, which was spun out of MIT’s Media Lab, opened up its software to SDK developers in 2014, but it’s now offering a free SDK and API to companies valued under a million dollars to spur creation. The software can only get better through more data, which makes outsourcing collection crucial to improving the deep learning of the technology.
“It’s kind of an infinite project, so we’re all about partners,” says Zijderveld.
Using increasingly intelligent neural networks, Affectiva can identify emotion across gender and age, Zijderveld says. The company’s next big challenge will be providing data that allows the technology to understand the situational contexts of emotions.
The company is also working on software that will recognize emotion in voices, opening up its potential in social communication and customer service.
Of course, a future where your emotions are always being monitored raises red flags regarding privacy. Affectiva requires all licensees to agree to terms that state they will ask for consent to turn on the camera to record someone’s face, Zijderveld says.
But as the market expands to other companies, concerns over data collection are sure to become a bigger issue.
“At the end of the day, we really believe this technology will be everywhere, but I also believe we’ll be okay with that,” says Zijderveld.
Photos via Affectiva, Youtube/Affectiva , Youtube/Affectiva