Problem with filters on social media

We tend to think of people we find attractive as intelligent, sociable, and trustworthy. This cognitive bias is known as the halo effect, which means we infer other characteristics from one characteristic of a person without any objective basis. If the distortion is positive, it is also called the “halo effect.” faint meaning “aura”), with the negative side effects of the “devil horn effect”.

The halo effect of physical attractiveness is well documented. The quicker we make a judgment about a person, the stronger this feeling becomes. The halo effect can lead to attractive people getting better grades or higher salaries.

The halo effect may also play a role in social media – many users now use beauty filters in their apps to optimize their faces according to common beauty ideals. Scientists have long worried that the popular trend of enhanced photos is upsetting to young women, particularly because it creates an unrealistic body ideal and makes real faces appear relatively unattractive.

However, whether and how the halo effect affects social media has been little studied to date. The non-profit research foundation Ellis (European Laboratory for Learning and Intelligent Systems) in Alicante has now shown that the halo effect does indeed work in beauty filters as well.

Test subjects rate faces

In research published by the Royal Society, a team led by computer scientist Nuria Oliver used a sample of 462 real facial photos and data from 2,748 test subjects to examine the effects of beauty filters on social perception and cognitive distortions. Subjects rated their “intuition” on randomly selected original faces and enhanced faces modified by popular beauty filter apps. They didn't know that filters were applied to some of the photos, and they didn't see the same person in either version.

These original photos (left) and filtered photos (right) were used in the study. Filters change skin tone, eyes and eyelashes, nose, chin, cheekbones and lips to make a person look more attractive. Image: royalsocietypublishing.org

Results: Test subjects clearly rated the people in the filtered photos as more attractive. On the 7-point scale used for scoring, the average increase was 1 point. Overall, test subjects rated no fewer than 96 percent of the photos modified with the filter as more attractive, and no single photo as less attractive.

Creates a halo effect even in photos with filters

So far, too. Test subjects not only rated the people in the altered photos as more attractive, but also rated them as smarter, more trustworthy, happier, and more sociable. In this regard, the research results also confirm the halo effect of beauty filters. However, there are exceptions, as co-author Aditya Gulati emphasizes: “Unlike attractiveness, not everyone was rated as smarter or more trustworthy, but in our study the statistics were still statistically significant for the entire sample. Increase.”

The gender and age of the person depicted affected perceptions – younger people and women were considered more attractive than older people and men, and women benefited more from the filter in terms of attractiveness. However, race did not affect the effectiveness of the filter and had no effect on perceived attractiveness.

Filters reinforce gender stereotypes

However, comparisons between female and male ratings proved illuminating: before and after applying the filter, photos of women had higher attractiveness ratings than photos of men. They also received higher scores on reliability, social skills and happiness, which is consistent with the halo effect. However, women had lower intelligence scores than men. «This is so surprising. “This suggests that gender stereotypes are masking the halo effect,” Gulati noted.

That's not all: the gap between men and women in intelligence assessments widened even further after using filters. Thus, filters appear to not only confirm existing gender biases but also reinforce them. Gulati added: “Equally important, beauty filters are primarily used by women, which makes this result even more concerning.”

Additionally, the effects of filters are not equally strong for everyone. Middle-aged people and women benefited more in terms of attractiveness, while men benefited more in terms of intelligence. It also showed that men and women were rated differently for attractiveness and other characteristics depending on whether the viewer of the photo was a man or a woman. Men rated their own attractiveness lower than women overall, but they were more susceptible to filters.

“Saturation Effect”

Of course, the effectiveness of beauty filters seems to be limited. As Gulati explains: “For some traits, especially intelligence, there appears to be a limit to how much a person is perceived to be smarter as attractiveness increases. An increase in attractiveness from a 5 to a 6 is the same as an increase in attractiveness from a 5 to a 6. There is a smaller increase in intelligence from a score of 2 to 3, but this is not the case for other traits such as social behavior. A similar “saturation effect” as for intelligence also exists for credibility. Perceiving.

This means that the correlation between attractiveness and intelligence becomes weaker for very attractive people. “Since filters increase the attractiveness of almost everyone, they could potentially be used to reduce the intensity of the halo effect,” Gulati speculates. However, Gulati explained that the research team does not support the use of beauty filters to reduce the halo effect. Their use can have many negative consequences, namely the reinforcement of gender stereotypes.

For the research team, the findings point to the need for regulatory guidance and public awareness regarding the use of such AI technologies in social media. (Human Resources Department)

The world's thinnest pasta? You might think we're in the realm of food. But that's not the case—ultrafine pasta, although made from flour like its thicker cousin, falls into the realm of nanotechnology. You can't eat pasta that's only 372 nanometers (1 nanometer = one billionth of a meter) thick because it will overcook so quickly that you can't remove it from the boiling water in time, as Gareth Williams of University College London explains explained.