A new psychology study suggests that the same selfie can change how “smart” and “good” we think someone is, once a beauty filter is added. Researchers tested how digital touch-ups on faces affect judgments of intelligence, trustworthiness and other traits that matter in daily life. Their results hint that beauty filters do more than smooth skin. They quietly shape who we see as capable, kind, or worth trusting.
This matters in a world where job recruiters, online daters and even teachers may all glance at profile photos. If filtered faces often “win,” then people who use these tools might gain subtle social advantages. At the same time, the study also found limits and unexpected twists, especially around gender and intelligence. So the story is not as simple as “prettier is always better.”
New research tests how beauty filters change first impressions
In the new research, a team led by Aditya Gulati, a PhD student at ELLIS Alicante in Spain, wanted to see if the classic “what is beautiful is good” effect still holds in the age of beauty apps. Their work, published in Royal Society Open Science, looked at how people rate faces that are digitally enhanced compared with the exact same faces in their original form. The study focused on fast, gut level impressions of strangers.
Psychologists have known for decades that more attractive people are often judged as more competent, honest and friendly. This pattern is called the attractiveness halo effect. Gulati and his colleagues asked a simple but modern question. Does that halo still appear when “attractiveness” comes from a filter, not from the person’s natural face?
For many of us, this question is not abstract. People use filters on dating apps, LinkedIn photos and casual selfies. A boosted image might change whether someone wants to hire you, date you, or trust you with a task. The researchers wanted data on how strong those shifts in perception really are.
How the study used original and filtered faces
To test this, the team gathered photos of 462 adult faces from two large image sets often used in psychology research. The faces covered a broad mix of ages, genders and ethnic backgrounds. All photos showed neutral expressions, so smiles or frowns would not drive the ratings.
Next, they ran each image through a popular mobile beauty filter. The filter smoothed the skin and also made subtler shifts, such as slightly larger eyes and fuller lips. The goal was to create a version of each person that many users would likely see as “better,” while still keeping the face recognizable.
Then came the ratings. The researchers recruited 2,748 adults, mostly from the United States and the United Kingdom, through an online platform. Each person saw only ten faces. Some participants saw only original photos. Others saw only filtered ones. No one ever saw both versions of the same person, which helped avoid obvious comparisons.
After looking at each face, participants rated how attractive, intelligent, trustworthy, sociable and happy the person seemed. These are traits that often guide quick decisions, like who seems safe, who seems competent, or who might make a good friend.
Filtered faces were rated as more attractive and more likable
The first big finding was not a surprise, but the scale of it was. In more than 96% of cases, the filtered faces were rated as more attractive than the originals. No filtered face was rated as less attractive. For a tool meant to boost looks, the filter clearly did its job.
The size of the boost, however, was not the same for everyone. People who started with lower attractiveness scores in their unedited photos saw the biggest jumps once the filter was applied. Faces that already scored high on attractiveness still improved, but the change was smaller. In other words, filters gave the largest “level up” to those who were initially rated as less attractive.
Those shifts did not stop with looks. Filtered faces were also seen as more trustworthy, more sociable and more happy. When a face moved up on attractiveness, ratings on these traits usually climbed along with it. The classic halo effect was alive and well in the digital world. A smoother, brighter selfie did not just look better. It also suggested a better person, at least in that quick moment of judgment.
At first glance, this might sound flattering. On a deeper level, it highlights how easily our brains tie someone’s character to their appearance. That can create quiet advantages for those who feel comfortable using filters and disadvantages for those who do not, or cannot.
Why the “attractive is good” effect weakens for intelligence
While the halo effect showed up for many traits, the pattern for perceived intelligence was more complicated. Attractive faces were still seen as smarter on average. Yet the link between looks and intelligence grew weaker at the very high end of attractiveness, especially for heavily beautified faces.
The researchers described this as a kind of “saturation.” If a face moved from low to medium attractiveness, intelligence scores rose quite a bit. Moving from medium to very high attractiveness gave a smaller bump in intelligence ratings. In plain terms, after a certain point, extra beauty did not buy much extra “brains” in the minds of viewers.
This is an important nuance. It suggests that people may see very polished faces as charming, happy, or sociable, but they do not always assume those faces belong to the smartest people in the room. Other traits benefit more from extreme attractiveness than intelligence does.
For anyone who worries that they need to look perfect to seem competent, this may be a small relief. The study hints that moderate improvements in appearance can shift perceived intelligence, but beyond that, the returns start to shrink. At some point, many viewers might treat a very filtered look as “pretty” more than “capable.”
Gender and age patterns in ratings of filtered faces
Gender and age added another layer to the story. Across the board, younger adult faces were rated as more attractive than older faces, both before and after filtering. However, beauty filters often did more to close the gap for middle aged and older people. The digital boost helped them catch up a bit in perceived attractiveness.
Gender patterns were even more striking. Female faces were usually rated as more attractive than male faces and also as more trustworthy, sociable and happy. So far, that fits the idea that attractive people get a general positive glow. But when it came to intelligence, female faces were often rated lower than male faces, even when they looked more attractive.
After filtering, the gap in perceived intelligence between male and female faces actually grew. The more polished and “beautiful” the women looked, the more their intelligence ratings trailed behind men’s ratings. This suggests that a gender stereotype was overriding the usual halo effect. In some viewers’ minds, being very attractive and being very intelligent did not seem to fit together as well for women as it did for men.
These patterns do not mean that women are less intelligent. They show how biases in the eye of the beholder can shape life chances. A woman who uses filters might gain points for warmth or likability, but she may also face a quiet penalty on perceived competence. For men, the trade off may play out differently.
What this could mean for selfies, social media and AI tools
On social platforms, many people use filters without thinking much about it. This study suggests those small choices might influence how others see their personality and abilities, not just their skin tone. Someone who habitually posts filtered photos could be seen as more sociable and trustworthy, especially if they start from a lower baseline of perceived attractiveness.
At the same time, heavy filtering may carry hidden social costs, which depend on gender and context. In a dating app, the boost in likability might feel helpful. In a work setting, a highly polished image could help or hurt perceived intelligence, based on the viewer’s own stereotypes. Different audiences may react in different ways and the study does not claim the same pattern will appear in every setting.
These findings also raise questions for artificial intelligence. Many AI systems work with images, such as face recognition or automated profile screening. If humans show a strong halo effect around filtered faces, AI models trained on human labels might learn that bias too. Over time, that could bake our beauty based assumptions into digital systems.
- Filtered faces may receive more positive trait labels in training data.
- AI models can learn to link beauty with trust or competence.
- Those links could affect decisions in hiring, dating, or moderation tools.
Consider: when you scroll through feeds filled with polished faces, your own mental “norm” shifts. Unfiltered faces can start to look tired or less capable, even if you know better. The study reminds us that our brains quietly connect looks with worth and that digital tools can push those connections further.
Limits of the study and where researchers want to go next
Like any study, this one has limits. The participants were mostly white adults from the United States and the United Kingdom and they all spoke English. That means the results may not match cultures with different beauty ideals or different stereotypes about gender and age. The faces also came from research image sets, not from real social media profiles.
The team used one popular beauty filter, not a wide range of apps and styles. Many filters exaggerate features more strongly or add makeup effects. Others are very subtle. It is possible that stronger or different filters would change the size of the halo effect, or even flip some results.
The researchers also focused on positive social traits like intelligence, trustworthiness, sociability and happiness. They did not ask about traits that might turn negative, such as vanity or shallowness. Future work could test whether extremely polished faces also seem more self centered or less genuine in some viewers’ eyes.
Looking ahead, Gulati and his colleagues plan to explore how AI systems pick up these human biases. They are already analyzing images from text to image models to see how those models “imagine” attractive people and what traits they attach to those faces. The long term goal is not to shame people who use filters, but to understand how human and machine judgments interact in a filtered world.