AI apps that rate facial attractiveness fuel looksmaxxing culture and the risk of body image disorders among young people

A growing number of websites and apps use AI to analyse a user's face and assign it an “attractiveness” score on a scale, potentially identifying supposed flaws in appearance or recommending ways to improve the score. As an example, the author describes a livestreamed video chat game in which AI compares two users and declares the one with the higher score the winner, while viewers ridicule the loser in the chat. According to the author, these tools pretend that attractiveness can be measured with scientific precision, even though beauty is subjective and varies across cultures, and AI systems also reflect biases in their training data – while a critical remark from a person can be recognised as an opinion, a precise number from a computer can seem like an objective fact.

These tools are part of the broader online culture known as looksmaxxing, in which users share advice on achieving a “more conventionally attractive” appearance – from grooming and exercise to promoting cosmetic procedures, extreme diets or dangerous practices. A 2026 review of looksmaxxing research, drawing on a limited body of academic and so-called grey literature, described, among other things, the practice of “bone smashing” (repeatedly hitting the face with hard objects in the belief that the injuries will narrow the jaw), for which the review found no evidence of effectiveness. The review also noted that high-quality research on looksmaxxing remains scarce, and recorded associations with constant comparison, body dissatisfaction and traits linked to body dysmorphia and eating disorders. According to a 2026 study examining young people's vulnerability to body image disorders, young men are particularly at risk of muscle dysmorphia – the belief that their body is not muscular enough.

The author also refers to her own research from the COVID pandemic, in which most participants with a history of eating disorders reported worsening symptoms during the British lockdown, in connection with disrupted routines, isolation and greater exposure to triggering content online. According to the text, eating disorders are associated with a significantly higher risk of premature death, which is one reason why new sources of pressure over appearance deserve detailed study – however, according to the author, the direct impact of AI ratings on user behaviour has not yet been scientifically demonstrated.

The text also highlights the potential misuse of AI appearance ratings within abusive relationships, because shaming someone over their appearance is a recognised feature of controlling behaviour, and a low AI score could reinforce the impression that an “objective” machine confirms a partner's criticism – however, according to the author, direct academic evidence for this specific misuse is lacking. It also mentions the risk that some apps presented as being for adults lack rigorous age verification, while services accessible to children in Britain are subject to the Online Safety Act, which imposes a duty to assess and reduce risks arising from content promoting eating disorders or bullying. Further details can be found in the source article.