This AI Can Learn Which Faces You Find Attractive by Looking at Your Brainwaves

The future might see you swiping right, but only in your brain.
photos people faces
Photo via Getty

Imagine a dating app that doesn’t require you to spend an exhausting evening just swiping in the hope of landing on a cutie. Now imagine this dating app playing matchmaker for humans in a way that it knows what kind of people you personally find attractive, giving you only those profiles to choose from. A new study makes this proposition sound like a very possible reality.

Researchers from the University of Helsinki and the University of Copenhagen have developed an AI that can learn which faces you find attractive by looking at your brainwaves.


The study published in February in IEEE Transactions in Affective Computing journal used 30 volunteers and monitored their brainwaves with the help of electroencephalography (EEG), which records the electrical activity of the brain. The team showed volunteers images stitched from 200,000 images of real celebrities and recorded their brainwaves responses. 

“It worked a bit like the dating app Tinder: the participants ‘swiped right’ when coming across an attractive face,” senior researcher and Docent Michiel Spapé from the Department of Psychology and Logopedics at the University of Helsinki said in a release on the study. “Here, however, they did not have to do anything but look at the images. We measured their immediate brain response to the images.”

Unlike dating apps, the volunteers didn't have to swipe right or left but the researchers could tell on the basis of EEG readings what faces they found most attractive. Researchers then analysed this EEG data using machine learning techniques generating a neural network. 

The data was then fed into an AI that created new images based on those preferences. The images, which were shown one at a time, were generated using a generative adversarial neural network (GAN) which created hundreds of artificial portraits. 


The researchers generated new portraits for each participant, predicting they would find them personally attractive. They found that the new images match the preference of the subjects with an accuracy of 80 percent.

“A brain-computer interface such as this is able to interpret users’ opinions on the attractiveness of a range of images. By interpreting their views, the AI model interpreting brain responses and the generative neural network modelling the face images can together produce an entirely new face image by combining what a particular person finds attractive,” academy research fellow and associate professor Tuukka Ruotsalo, who led this project, said.

Researchers believe this technique can also be used to study unconscious attitudes and preferences like race and ethnicity, which make a person attractive and could mean computers start understanding human preferences in the future. “If this is possible in something that is as personal and subjective as attractiveness, we may also be able to look into other cognitive functions such as perception and decision-making. Potentially, we might gear the device towards identifying stereotypes or implicit bias and better understand individual differences,” said Spapé in a statement. 

Spapé further explained that in previous models, people usually found models who are blond and smile attractive. In a YouTube video released by the researchers, the individually generated photos feature only white people. 

In 2016, an AI-judged beauty contest led to nearly all-white winners. The contest saw 600,000 entrants, who sent in selfies from around the world—India, China, all over Africa, and the U.S. They let a set of three algorithms judge them based on their face’s symmetry, their wrinkles, and how young or old they looked for their age. The algorithms did not evaluate skin colour. The results were shocking: Out of the 44 people that the algorithms judged to be the most “attractive”, all of the finalists were white except for six who were Asian. Only one finalist had visibly dark skin. The problem, in such cases then, is with the lack of diversity of people and opinions in the databases used to train AI, which are created by humans. In the new study however, the AI is getting trained on personal preferences, which is where it gets even murkier. It then seems that Eurocentric ideals of beauty are rooted in our psyche way more than we’d imagine.

“Attractiveness is a more challenging subject of study, as it is associated with cultural and psychological factors that likely play unconscious roles in our individual preferences,” Spapé said at the end of the statement. “Indeed, we often find it very hard to explain what it is exactly that makes something, or someone, beautiful: Beauty is in the eye of the beholder.”

Follow Jaishree on Twitter and Instagram.