AI Face: plastic surgeons are increasingly worried about patients who want to look like their AI-generated versions
Unrealistic expectations have always been part of the journey for anyone considering plastic surgery. Magazine references, celebrities, and social media filters have been giving surgeons headaches in their offices for a long time. But what is happening now is different, and specialists are genuinely concerned.
More and more patients are showing up at consultations with something new in hand: images of their own face transformed by artificial intelligence, ready to serve as a reference for the surgeon. Flawless skin, sculpted cheekbones, a refined nose, near-perfect symmetry. The problem is that these images are not a surgical blueprint. They are pixels. And the difference between manipulating pixels and manipulating human anatomy is enormous, as we will explore throughout this article.
British plastic surgeons are already reporting a significant increase in this behavior in their offices, and the phenomenon has even earned a name: AI Face. It is the idealized version of yourself, created by a chatbot, that will probably never exist outside of a screen.
Dr. Nora Nugent, a cosmetic surgeon based in Tunbridge Wells and president of the British Association of Aesthetic Plastic Surgeons, has been seeing this firsthand. According to her, patients are showing up with AI-enhanced photos of themselves and the false expectation that those results are achievable through surgery. And many of her colleagues report similar experiences.
I can only foresee an increase, given the speed at which AI has been incorporated into all aspects of life, Nugent said.
Understanding why this has become a serious problem, both for those who want the procedure and for those who perform it, is exactly what we are going to explore here. 👇
What AI-generated images are and why they look so real
The AI-generated images showing up in consultations are not simple Instagram filters. They are produced by advanced visual language models, like Midjourney, DALL-E, or specific facial transformation tools, which analyze facial structure and recalculate every detail based on patterns of ideal beauty extracted from millions of images. The result is something that looks photographic, hyper-realistic, and technically possible, but in practice completely ignores how human anatomy works. There is no bone, cartilage, muscle, or skin in those images. There is only what the algorithm understood as visually pleasing according to a generic aesthetic standard.
The real danger lies precisely in that appearance of reality. When a person sees their own face transformed at that level of detail, the brain tends to process it as a concrete possibility, not as a computational illustration. The feeling is that all it takes is a consultation and a few weeks of recovery for that version of the face to become real.
This cognitive leap is the starting point for a series of unrealistic expectations that put both the patient and the surgeon in an extremely delicate position. One is convinced the result is achievable, and the other knows it is not.
Dr. Alex Karidis, a surgeon based in west London, put it quite bluntly: while AI can control each pixel individually, surgery certainly does not work at that microscopic level of detail. The pixel-by-pixel precision that a generative model offers simply has no equivalent in the world of medicine.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
This phenomenon did not come out of nowhere. It is the natural evolution of a behavior that already existed with social media filters, but gained a new layer of sophistication with the widespread adoption of generative AI tools. Before, filters were visibly artificial. Today, images generated by AI models have a quality that challenges human visual perception, making it much harder for users themselves to distinguish what is real from what was created by a computational system trained to produce images of high aesthetic fidelity.
AI Face: the aesthetic standard that plastic surgery cannot deliver
The term AI Face describes a specific set of characteristics that appear repeatedly in images generated by artificial intelligence when the prompt involves human faces considered attractive. These are traits like near-mathematical facial symmetry, skin completely free of visible pores, a defined jawline without any texture variation, and a nose with precise tips that rarely exist in natural anatomy.
Surgeons have already noticed these consistencies in what patients bring as references, and hyper-symmetry is the most frequent issue. AI can generate perfect symmetry effortlessly, but recreating that in a real human face is often impossible.
Dr. Julian de Silva, a cosmetic surgeon on Harley Street in London, illustrates the problem well. If one of your eyes sits a few millimeters higher than the other, AI corrects it in seconds. But rearranging pixels is not the same thing as rearranging anatomy. De Silva explains that it is impossible to change the position of the eyes because that is set in bone, and the brain sits behind the orbits. You cannot safely change the position of the orbits.
De Silva also observed that when AI edits a patient’s photo, it frequently defaults to widely accepted beauty standards: for women, a V-shaped jawline, an ogee curve along the cheekbones, and a heart-shaped face. For men, wider jaws, lower brows, and fuller upper eyelids.
The problem with the aesthetic standards imposed by these images goes beyond the technical question. They also carry an enormous cultural bias, since AI models are trained on datasets that historically overrepresent certain Western aesthetic standards and underrepresent the real diversity of human faces around the world. The ideal beauty that AI projects onto a person’s face is neither neutral nor universal. It is a very specific slice that was repeated so many times during model training that it became the dominant standard in generated images.
Plastic surgeons report that the difficulty is not just in explaining the technical limitations of the procedure, but in deconstructing the emotional belief the patient has developed around that image. Both Karidis and Nugent describe how AI-generated images are psychologically effective at setting and reinforcing patients’ aesthetic ideals.
Nugent is emphatic: once you see an image, it hooks into you. Karidis agrees, describing AI images as something that gets branded into patients’ minds, and said that colleagues have recently been flooded with this type of reference.
The surgical consultation stops being a technical planning session and becomes a negotiation about expectations, and that negotiation often ends with frustration on both sides because the starting premises are simply incompatible.
A journalist’s experiment in the surgeon’s office
To better understand the phenomenon in practice, The Guardian journalist Isaaq Tomkins decided to run a test. He asked an AI agent to recommend cosmetic procedures and generate images for surgeon Alex Karidis to evaluate. The results are revealing.
In the first round, the chatbot suggested a rhinoplasty and septoplasty, refining the nasal tip and straightening the bridge. It also applied a subtle blepharoplasty, which is an eyelid lift, and brow refinement. Karidis assessed that the rhinoplasty was relatively modest and the blepharoplasty nearly imperceptible, but estimated the work would cost around 25,000 pounds.
In the second round, things started to escalate. Tomkins asked the virtual assistant to give him predator eyes and a more masculine face. The AI recommended chin implants, buccal fat removal, infraorbital augmentation, another blepharoplasty, facial beard grafts, and a series of other procedures.
Karidis’s reaction was blunt: this is where things start to get a bit ridiculous. It looks like they gave you someone else’s eyes. He said the chin implant was unnecessary and that buccal fat removal would take its toll later in life, as the face naturally becomes leaner with age.
If someone were to theoretically undergo everything the AI suggested, the cost would easily exceed 100,000 pounds and would still probably look nothing like the generated image, not to mention the exposure to potentially significant side effects and recovery time.
In the third and final round, Tomkins asked the chatbot: make me look more chad. The AI responded with additional recommendations, including a neck lift, brow lift, two types of custom implants, and full ablative laser resurfacing to create perfectly even, renewed skin.
Karidis’s final assessment was damning: this is where things start to look scary. What are those enormous holes along the jaw angle? It looks like chunks of tissue have been removed. As for the neck lift and brow lift, that is frankly bogus. I see no evidence of lifting in those areas. Tissues like the brows appear to have been lowered rather than lifted. Your original skin looks much better than this.
This experiment clearly illustrates a point that surgeons are keen to emphasize: no matter how much patients fixate on the visual results generated by AI, the real results of surgery are subject to physical, biological, and financial limitations that no algorithm can work around.
The problem with fake results on social media
Beyond patients who arrive with AI references, there is another growing concern among surgeons: clinicians who share surgical results on social media that look spectacularly effective but may themselves be generated or manipulated by AI.
Dr. Julian de Silva reported seeing a case that caught his attention the week before his interview. It was a video in which a patient appeared to have rejuvenated by 30 years after a procedure. De Silva watched the video multiple times, intrigued by the impossible result. On the third viewing, he noticed the telltale detail: the hands had six fingers.
This type of content is particularly dangerous because it comes from sources the public tends to trust, such as medical professionals. When a clinician publishes digitally manipulated results, they are not only misleading potential patients about what is achievable, but also eroding trust in the legitimate work of other professionals who present real results, with all the nuances and imperfections that reality includes.
The real impact on mental health for those seeking plastic surgery
The relationship between plastic surgery and mental health has been a field of study for decades, and what researchers already know is that patients with unrealistic expectations about surgical outcomes have much higher rates of post-operative dissatisfaction, regardless of the technical quality of the procedure performed. When those expectations are fueled by AI-generated images showing a computationally perfect version of their own face, the risk of dissatisfaction grows significantly, because the benchmark the patient carries in memory is literally unachievable by medicine.
Nugent makes a point of being clear about this with her patients before any procedure: the patient needs to understand that there is human variation in how you heal, how you age, and in what can be done. I tell patients upfront: what I can do in surgery is not unlimited. None of us control everything.
Mental health specialists who work alongside plastic surgery teams are already developing specific protocols to identify patients who arrive with AI references. The main red flag is the rigidity of expectations: patients who come in with AI-generated images tend to have less flexibility to accept that the outcome will differ from the reference, and that inflexibility is an important clinical signal that needs to be evaluated before any procedure. In some cases, the recommendation is to postpone surgery and begin psychological support specifically focused on body image perception.
There is also a cycle that deeply concerns professionals in the field: a person undergoes surgery based on expectations fueled by AI images, the result does not match the idealized digital version, the dissatisfaction leads to a new search for procedures, and the process repeats. This pattern is directly related to body dysmorphic disorder, a condition in which a person develops a distorted and obsessive perception of aspects of their own appearance. AI, in this context, does not cause the disorder, but can act as a powerful catalyst for anyone who already has a predisposition toward this type of thinking.
What healthcare professionals are doing about it
The medical community’s response to this phenomenon is still taking shape, but some guidelines are already emerging. Plastic surgery associations in countries like the United Kingdom, the United States, and Australia are discussing the need to include specific questions about AI tool usage during initial patient screening. The goal is not to create a barrier to accessing procedures, but to ensure that the informed consent process is genuinely informed, which means the patient needs to clearly understand the difference between what an AI tool can create visually and what plastic surgery can deliver anatomically.
Some practices have already adopted the use of surgical simulation software during consultations, specifically to replace AI references with projections that respect the real limits of the patient’s anatomy. These software tools are quite different from generative AI tools: they work with clinical photographs and medical parameters to show possibilities within what is technically feasible, without creating expectations based on ideal beauty standards disconnected from biological reality. Patient reception of this approach has been positive when the professional dedicates enough time to explain the difference between the two forms of visualization.
Karidis also warns about something many patients underestimate: when they do their deep research on cosmetic procedures, they tend to fixate on the images and ignore all the noise around them. The disclaimer that the chatbot itself eventually makes about the viability of the operations is simply ignored. For Karidis, this is the central point for everyone: the moment you show something like that to someone, that is it. It is over.
Public education also emerges as an important front in this debate. The more people understand how generative AI models work and what biases are embedded in the aesthetic standards they reproduce, the more critical they will be when using these tools as a reference for health-related decisions. This is not about demonizing the technology, which has genuinely useful applications in medicine and other fields, but about developing a visual literacy that allows users to recognize what they are seeing and what that image actually represents. 🤖
Artificial intelligence is a powerful tool, but when it starts defining what you think your face should look like, it is worth taking a step back and remembering that the algorithm will never see you in the mirror in the morning.
The debate about the boundaries between technology, aesthetics, and mental health is only just beginning, and the decisions that medicine and society make now will define how this intersection develops in the coming years. What is already clear is that pixels and scalpels play on completely different fields, and mixing the two without the necessary care has real consequences for real people. 💡
