AI told me exactly what I needed done to my face - should I trust it?

Settling down on a treatment bed at Harley Street’s The Hale Clinic, my face was assessed, measured and analysed before the aesthetics practitioner Dr Ivona Igerc (known to me as Dr Ivy) went anywhere near me with numbing cream, let alone a needle.

Before deciding where to inject filler and neuromodulators, she mapped out my face on her iPad, marking which areas she would inject and what product she would use. "It's such a precise art," I said, so I was shocked when Ivona told me there has been a huge rise globally in robotics injecting facial fillers and neuromodulators. 

There are reported benefits of having a machine inject your face rather than a person, according to the National Library of Medicine, which highlights increased precision, accuracy and consistency of results. The cons? The lack of personal touch, limited experience and risk of injury due to malfunction. Not worth it, if you ask me.

When aesthetics meets AI

That's not the only way artificial intelligence is interfering in the world of aesthetics, though. People are increasingly turning to ChatGPT for aesthetics advice, uploading photos of themselves to the platform and receiving a breakdown of which treatments to book within seconds, as well as images generated on the platform of how they might look post-procedure.

This is something Mr Samit Unadkat, a consultant ENT surgeon in the NHS and at My Nose London, has seen a rise in, telling hello!: "One in ten rhinoplasty patients now arrives at consultation with an AI-generated image of their own face."

I uploaded a photo of myself to ChatGPT to see what it would suggest I needed. My robot recommended neuromodulators, advising: "If you want to soften or prevent the fine lines that naturally appear when we express ourselves – like faint lines on the forehead or around the eyes (crow's feet) – a very subtle touch of a neuromodulator like Botox, Dysport or Xeomin is the standard approach."

© Phill Taylor

When it came to skin texture and radiance, it gave me a few compliments first. "You already have a lovely complexion, but if you want to boost collagen, even out skin tone and maximize that glow, you could look into microneedling – great for overall skin tightening, boosting collagen production and refining texture."

Not too worrying, and perhaps useful as a jumping-off point, but can AI recommendations be trusted? Dr Kaywaan Khan, Private GP and medical aesthetics doctor at Harley Street’s Hannah London, thinks not.

"AI can be a genuinely useful starting point when it comes to skin analysis. Through uploading a simple selfie, it can pick up on visible features such as redness, blemishes or uneven tone, giving someone a first impression of their skin before they've even booked a consultation."

© Getty Images

That's where the uses end, though, according to Dr Khan. "In-person consultations remain the gold standard, and that's not about the human eye outperforming advanced technology. It's about everything a medically trained clinician does that a photo simply can't offer. They'll feel how the skin responds to touch, check whether an area is raised, and look at it from angles a single photo would never capture. However advanced the AI reading it is, the image itself is still the limiting factor."

Mr Unadkat agrees: "An image cannot account for the way a change to one feature shifts the relationship between every other feature on the face."

Empty compliments

The fact that AI is programmed to tell us what we want to hear is an issue, too. My ChatGPT consultation hyped me up before diving into the suggested treatments. "First things first: you have gorgeous, healthy-looking skin with a fantastic natural glow, bright eyes and great bone structure. You absolutely do not need any aesthetic treatments," it told me.

Empty praise aside, the suggestibility of AI is one of the issues with such consultations, says Dr Khan. "Ask ChatGPT whether vitamin C would suit you, and it's likely to respond with the general, well-documented benefits of the ingredient, without first asking whether you've used it before or reacted badly to it."

In contrast, a proper assessment with an aesthetics professional starts with those kinds of questions, not the other way around, which is why AI-generated advice should always be taken with a pinch of salt rather than followed blindly, Dr Khan says.

He adds: "AI is built to sound confident, whether or not the answer it's giving is complete. That confidence can make it hard to tell the difference between general information and advice that's genuinely personalised, especially when the response is detailed and reads as though it understands your skin specifically."

© Getty Images/Maskot

Unrealistic expectations

If someone goes so far as to create an AI-generated image of how they hope to look post-procedure, they can find themselves with unrealistic expectations of what they can achieve, according to Mr Unadkat.

"When a patient brings in a photograph of someone else, there is an implicit understanding that this is a different person," Mr Unadkat says. "When they bring in an image of their own face, the psychological gap between expectation and reality is much harder to bridge."

Indeed, a peer-reviewed study published in Aesthetic Plastic Surgery found that patients with prior experience of AI image enhancement held significantly higher expectations for surgical outcomes – and are thus predisposed to lower satisfaction after surgery.

The bottom line? Go ahead and put your photo into ChatGPT for initial advice, but don't take it as the final word. In the same way we can't blindly book in for the exact procedures celebrities have and expect the same results, we need personalised evaluation to ascertain whether a procedure will be right for us.

"AI can be genuinely useful for helping someone understand their skin a little better before an appointment; however, the diagnosis and the treatment that follows need a qualified medical professional behind it," Dr Khan says.