Noah Levenson’s interactive film shows the broken promises of facial emotional recognition AI

It feels like 2019 is the year where AI has finally seeped seamlessly into day-to-day life. Yes, there is still a sense of novelty to it and the technology is still far from the utopian dreams of ubiquitous computing, but it seems like everywhere you turn, everything suddenly seems to be powered by AI. From suggesting what to write in your emails, helping type designers create new fonts to giving you feedback about how to brush your teeth better, this technology seems to be more visible than ever.

But it seems like the more you see, the less you know, and this is precisely what filmmaker and programmer Noah Levenson shows us with his seamlessly interactive film, Stealing Ur Feelings. The six minute film runs in your browser and deals with the topic of facial emotion recognition AI that’s being patented and used by the likes of Facebook, Snapchat and Google all the time. It’s rare to find an interactive film that’s so well-crafted – from the humour, editing, technology, and the general sense of bewilderment that it creates. The Brooklyn-based former MTV creative executive does this by showing us not what we see, but what the machine sees – all without a sniff of the damp padding of a pair of overworn VR goggles.

“I’m mostly interested in trying to make things that nobody has tried to make before. Sometimes that means trying to mutate the medium of film by merging it with computer science,” Noah tells It’s Nice That. Growing up in a working class suburb of New York, Noah wants his output to be consumable by everyone. “I like to make simple entertaining things that everyone can appreciate without reading an artist’s statement,” he continues.

The film, which is made in the spirit of overenthusiastic educational YouTubers, shows you different clips by using actual emotion recognition AI to measure how positive or negative your reactions are to what they show you on screen. The premise is that every time you calibrate your face for facial recognition anywhere – usually in the form of face filters – the algorithm has the ability to continually calculate your emotional response based on the different ratios on your facial features. “It’s so fast and invisible that you don’t know that it’s learning all kinds of interesting things about you,” the narrator eagerly explains.

“Depending on your reactions, the film makes a series of decisions about you – in ways that you probably wouldn’t want a machine making decisions about you,” Noah says. Feeding off this data, watching you watch the film, the film shows you the proportion of men and women that might appear on your dating apps, the type of news that might be shown to you, and if you appear to have a racial bias. “Welcome to your new life as an input!” the narrator announces.

If it sounds ridiculous that an algorithm might make decisions about you based on whether you smiled or frowned at a picture, you’re right and Noah’s goal is to show you how easily deployed these emotional recognition technologies are. “I hope viewers learn a little bit about the current state of artificial intelligence and the ways that their favourite apps could be using them to make their lives, and the world, worse,” Noah says. Surely enough, the film shows that these companies have quietly patented their plans for what they aim to use this technology for.

Despite his stance on the usage of this technology, Noah is far from a technophobe, capturing the spirit of the early internet in his enthusiasm for computing. “It was the 80s and we were the only people in our neighbourhood who had a computer. We had just moved there from artists’ housing in Manhattan, and me and my brother were total outcast weirdo nerds,” Noah describes. “So I hid inside and spent all my time on that computer, first learning Basic, and then Pascal and C.”

He is, in fact, excited that computers are everywhere. Perhaps the multi-layered nature of computing, where each level of abstraction “reveals a deeper, truer level of reality", is why he so eagerly tries to demystify facial emotional recognition AI. It is engaging no matter what your level of technical knowledge is. “The field is currently dominated by people from fancy academic institutions and the traditionally exclusive film and media structures,” Noah says. “Everyone is morally posturing about how much they care about society, but no one has actually talked to a poor person before.” An educational gem that’s just so hard to stop watching, the film is a masterclass on how to create interactive media that is both critical and enjoyable at the same time.

Watch the film at www.stealingurfeelin.gs