The latest technology and digital news on the web

Human-centric AI news and analysis

Can AI read your emotions? Try it for yourself

Emotion acceptance AI is bunk.

Don’t get me wrong, AI that recognizes human affect and affect be very useful. For example, it can help analyze when drivers are falling asleep behind the wheel. But what it cannot do, is anticipate how a human being is absolutely activity by the announcement on their face.

You don’t have to take my word for it, you can try it yourself here.

Dovetail Labs, a accurate analysis and consultancy company, afresh created a website that explains how modern “emotion recognition” systems built on deep acquirements work.

Typically when companies do stuff like this, the point is to show off their articles so you’ll want to acquirement commodity from them. But here, Dovetail Labs is demonstrating how awful affect acceptance is at acquainted human emotion.

If you’re a bit bashful to enable your webcam for access (or just don’t feel like trying it out on yourself), take a gander at the featured image for this commodity above. I assure you, the picture on the right is not my “sad” face, no matter what the AI says.

And that’s aggravating because, as far as AI is concerned, I have a pretty easy to read face. But, as Dovetail Labs explains in the above video, AI doesn’t absolutely read our face.

Instead of compassionate the vast spectrum of human affect and expression, it basically reduces whatever it perceives we’re doing with our face to the AI agnate of an emoji symbol.

And, admitting it being abundantly basic, it still suffers from the same biases as all facial acceptance AI: affect acceptance systems are racist.

Per the video:

A recent study has shown that these systems read the faces of Black men as more angry than the faces of white men, no matter what their expression.

This is a big deal for everyone. Companies across the world use affect acceptance systems for hiring, law administration agencies use them to contour abeyant threats, and they’re even being developed for medical purposes.

Unless affect acceptance works the same across demographics, their use is adverse – even in alleged “human in the loop” scenarios.

Published April 5, 2021 — 21:57 UTC

Hottest related news

No articles found on this category.