Will the robot doctor see you soon?

As discussions around artificial intelligence continue to grow, contributing reporter Becki Murray explores how it might positively and negatively affect the aesthetics industry over the next few years

Picture the scene. A patient walks into your clinic and, on entry, their face is scanned in intricate detail to help recommend the best procedures and at-home skincare to tackle their skin goals. Post-consultation, they enter the treatment room where a robot delivers exactly what they need, be it body contouring, injectables, or even cosmetic facial surgery. According to artificial intelligence advocates, this could be the image of your modern clinic in the next few decades. Yet, while technological advancements are making greater precision seem closer than ever, there are real concerns about the safety and potentially biased data behind these AI advancements.

WHAT IS ARTIFICIAL INTELLIGENCE?

Firstly, a recap. Artificial intelligence, AI for short, is technology designed to ‘mimic’ human decision-making. Algorithms play a significant role in this ‘machine learning’, working as a set of rules input into a computer programme, which helps dictate appropriate responses. (That’s why AI is only as good as the data you put in).

Indeed, AI is probably influencing your life right now. From streaming platforms like Netflix, to social media, and at-home assistants such as Alexa, our digital habits provide a ‘footprint’ that AI uses to provide personalised recommendations and advertising. Even the medical profession is embracing it. In 2019, the NHS launched an AI laboratory to fuel research into AI-based screening technologies for the health service. In 2021, it received £36 million in additional funding.

POSITIVE USES IN AESTHETICS

Ultimately, the biggest strength of AI in aesthetics is the tailored information it can provide for both practitioners and their clients. “AI has the potential to revolutionise the way we approach treatments,” says Charles Mills, chief executive officer of British wellness and aesthetics company, Gelida. “From personalised skin recommendations to real-time analysis gathering data on a client’s skin type, concerns and goals, we can use that information for a more tailored treatment approach, while still prioritising the importance of the human touch in all our devices.”

The cosmetic surgeon Dr Riccardo Frati agrees: “The technology could be applied and developed for breast, body and facial cosmetic surgery assessment, facial mapping, skin analysis and more custom made and tailored treatments plans.”

Haut.AI, for example, uses a database of millions of skin images to power its Skin SaaS algorithm-based face evaluation software. It looks at 15 skin metrics, which practitioners can use to create personalised skincare routines for clients. The latest version of the VISIA Complexion Analysis System also simulates how a patient’s skin could age without intervention, helping guide treatment plans. As AI advances, there are opportunities to visualise the results of injectable cosmetic treatments too. “By using AI technology that scans a patient’s face, and then allows you to alter their image, you can show a client what a little bit of cheek or lip filler could look like,” explains Dr Asha Chhaya, the senior medical injector for Sk:n Clinics.

Indeed, the Swiss pharmaceutical company Galderma has just launched an augmented reality (AR) tool that provides a visual simulation of the results of injectable dermal filler treatments such as Restylane and collagen stimulators like Sculptra (although it’s currently not available in the UK). Developed in partnership with Crisalix, a world leader in augmented reality aesthetic simulation, it allows practitioners to evaluate a patient’s face, including skin quality, facial shape, symmetry, and expression, giving both practitioners and patients confidence as they move through a treatment plan. There are even apps entering the space. Aedit is a 3D-aesthetic simulator that allows individuals to ‘try on’ procedures such as dermal fillers and rhinoplasty at-home – like virtual before-and-after photographs.

AI ROBOTICS

Then, there are the robotic surgery tools, which promise repeatable precision and accuracy. In hair transplantation, developments in robot-led follicular unit extraction mean an otherwise lengthy, repetitive process can be conducted autonomously by a robotic arm; an example is the FDA-approved ARTAS system, which extracts and implants (currently only dark, straight) hair, which the practitioner oversees. In laser therapy, EON by Dominion Aesthetics uses robotic arms to deliver touch-free ‘body mapping’ and FDAapproved abdominal fat reduction; the Italian company ASA has its hands-free M8 well-being laser and other AI-powered bodyscanning technology for adaptable pain relief rehabilitation, and the Slovakian company Saharra utilises a semi-autonomous robot assistant for its laser hair removal device to maximise precision and speed. Even the concept of AI-led robot injectors is increasingly possible. Parisian tech company NextMotion has introduced LENA, the ‘Light Enabled Neuro-Robotic Arm’, which can maintain sub-millimetre precision without tiring, although the doctor must still specify the quantity and location of injection.

Finally, “AI could help with planning for surgical interventions by predicting outcomes,” says Dr Pundrique Sharma, a cosmetic surgeon for Harley Medical Group. “At the hospital I work at, we deal with children with craniofacial, and chest wall anomalies or asymmetrical breasts, and you can use 3D scans, artificial intelligence-based programmes, and algorithms to help work out where differences can be mitigated for the best results. That’s something AI is good at – looking for trends to see what is and isn’t working. Hypothetically, you could also envisage AI models sifting through data to see whether certain implants cause more complications, whether certain procedures are safer, and which operations make which people happier.”

UNEASE ABOUT AI

However, there is plenty of apprehension about AI use. Firstly, considering AI relies on collecting your data and digital ‘footprint’ to personalise recommendations, there are privacy concerns that can cause patient uneasiness. Practitioners must also follow stringent GDPR data protection guidelines to avoid monetary and reputation-affecting breaches. Furthermore, when I ran “What are the limitations of AI in aesthetics” through a search engine AI tool, the response was extremely limited. Silencing my rather science-fiction-generated thoughts about machines ‘protecting’ one another, it emphasised another considerable drawback of the technology – the incomplete data it currently runs on.

Consider the algorithms that power our digital searches: “I don’t think AI algorithms are inherently evil, but they tend to be written in a way that maximises hits, with a focus on what generates the most money,” says Dr Sharma. “It’s those factors that determine what turns up at the top of your Google search, meaning there is a danger that the choices you are encouraged to make are not necessarily the best, healthiest, safest or morally appropriate choices for you, but the ones that are most lucrative for someone else.”

“It’s also causing an artificial idea of what normality is because the AI algorithm is focused on what most people are clicking on, not what is most accurate,” continues Dr Sharma. “That’s a massive concern because when all you see is things that are unrealistic, it distorts your opinion of what ‘normal’ is.” That’s not something we should blindly bring into the clinical environment. Likewise, because humans program AI, we influence it with our inherent biases. “The aesthetics industry isn’t currently always as data-driven as you may think; the data we are driven by isn’t always great; and if the data we plumb into an algorithm isn’t accurate, the results we get out won’t be either,” reveals Dr Sharma. “Moreover, while there are diverse cultural norms around the world, the internet is very Western-dominated, and AI programming is very male-dominated. It’s not that the AI is therefore racist or sexist, but we all carry certain, often unconscious, biases. That means there’s a risk of that bias being transferred to the AI used to produce the content you look at, which itself influences the choices you make about how you should look.”

“I do think AI models could be very useful, but we need a lot more data, and specifically diverse data because the modelling is based on a narrow perception of beauty,” agrees Dr Chhaya. “For example, we might have data on what’s considered ‘masculine’ and what’s ‘feminine’, but what if we have someone come in who doesn’t fall into those binaries?”

Robotic tools can face similar difficulties: “Robot injecting could be beneficial if we can replicate precise results accurately, but patients have different muscle size and depth, as well as different preferences.” It also comes down to knowing your clientele. Some potential patients may continue searching for the very latest advancements, yet many are more concerned about establishing trust.

At least for now then, experienced practitioners still trump technology when it comes to ensuring the personalisation and safety required for natural-looking results. Likewise, while AI-based recommendations could make the practitioner’s life easier when prescribing skincare, you don’t want to remove the expertise and reassurance of practitioners. “There’s so much more nuance to the face than what that kind of modelling can show,” says Dr Chhaya. “If a tool like that is going to be used, it needs to be part of a consultation, because you really don’t want to lose the human touch. We need to take the power of AI as a tool – and combine it with the skill of the practitioner.”

Furthermore, while AI use is only likely to accelerate in society, “it absolutely needs to be regulated,” before it reaches mass usage in aesthetics, concludes Dr Sharma. By campaigning for that, aesthetic doctors can help ensure the safety and accuracy of its use.

Until then, it’s wise for any aesthetic practitioner to see AI as a supporting and still developing tool in a much wider arsenal of their expertise. Expertise, you may be relieved to hear, that still currently comes out on top against the robots.