Companies are trying to automate the diagnosis of skin conditions like Acne and wrinkles with apps like Olay’s Skin Advisor and TroveSkin (which promises to “ your pores, wrinkles and spots” for “personalized skincare”). But how effective are these skin care apps? And, should you rely on them to find a skincare solution? Let’s find out.
How AI Skin Care Apps Work
Most of these skin care apps are straightforward and easy to use. You generally have to upload a selfie, answer some questions about your skin like – how you maintain it or what bothers you most about it – and then they diagnose your condition and recommend a solution like buying an anti-aging cream (for $22) to fix the lines around your mouth or something similar.
If you try another app out of curiosity, you will most likely get a different recommendation like – improve the texture of your skin using exfoliating products and brighteners. Additional benefits include tracking the condition of your skin over time and finding out the age of your skin. Yes, skin age is a thing. And yes, you should be afraid. Be very afraid! It’s great when the app underestimates the age of your skin But, if it overestimates, then the app makers have a serious problem. 😛
Demand for Skin Care Apps
The total revenue from cosmetics is expected to reach $ 805 billion within the next 5 years. And, skin care will account for much of it. With so many beauty stores, dermatological clinics and online retailers wanting to improve your skin, it can be difficult deciding whom to rely on. Plus, there are so many possible problems! Blackheads, redness, ageing. All you want is to make your skin look “dewy”, I know. Don’t ask me what that means.
There are several apps using artificial intelligence to diagnose skin conditions and making product recommendations based on inputs from the user. Neutrogena’s 2018 release – Skin360 – requires you to use a scanner (read 30x zoom lens) for a super close look at pores, fine lines and moisture level of your skin. Proven, on the other hand, draws on medical research published in journals to identify your skin condition and suitable product solutions.
Goal Is Personalization
Personalization is a big trend in products that use artificial intelligence technology. Consumer products like Netflix and Google Home have already shown how effective and popular they can be because of it. It was only about time, then, that skin care companies got on the personalization band wagon.
These technologies all work in the same manner. They collect your data, estimate your needs and match this information with product features to find the best pick. Easy, right? Furthermore, if they prove to be effective, you will have personalized skin care recommendations in real time for perfect skin all day every day. Personalization is the future of all products and services.
Proven’s co-founder Ming Zhao got the idea for the AI-powered beauty company after struggling to find a skin solution herself. She tried several “miracle” skin care products while working 16-hour shifts and found all of them lacking or to be ineffective altogether. It was only after she met her skin guru, who provided a personalized plan for her, that she found a solution. She recognized the need for personalized skin care services and turned to technology to provide it.
Her co-founder at Proven – Amy Yuan – is a computational physicist, who studied at Stanford. Together, the duo is building a database of skin care information using deep learning technology. Their database covers over 8 million customer reviews, 4 thousand journals, 100 thousand products and 20 thousand ingredients. Proven asks questions about user’s reaction to products, their sleep, etc., before their machine algorithms start to find the perfect solution. Zhao and Yuan even collaborated with a dermatologist to ask the right questions and offer effective solutions.
Good Quality Data Is Key
The main issue with skin care apps is quality of data that they collect. More often than not, it tends to be of low-quality – a blurry image, or inaccurate answer to a question. This makes it very difficult for the computer to provide an effective solution.
Consider one user’s experience. When she used her laptop camera, Olay’s skin advisor estimated the age of her skin to be 23. When she used her phone camera, the app thought her skin was 35 years old. Her actual age was 29!
Clearly, the quality of the image had a bearing on the computer’s estimate of the skin age. It’s not necessary for her skin to have the same age as her actual age, mind you. Your skin age can be different from your actual age. But the 12-year difference in the age estimates of the same skin shows the computer is highly dependent on the quality of data it receives.
Upload a selfie, log in information like how you are feeling, how much sleep you have had, how much you’re exercising and what you’re eating, and do all this daily. Skin care apps will love you for this. They will even offer you discounts on possible solutions. It’s the kind of data that will help them improve their recommendations over time.
What do dermatologists have to say about this method of diagnosis? One dermatologist was highly skeptical. She thinks that skin care apps have no clinical skill whatsoever and that they that are programmed to spit out a one-size-fits-all solution using algorithms.
She also has problems with the questions that skin care apps like Proven ask and thinks customers may find it difficult to answer questions about the level of stress they’re experiencing and other subjective questions. Finally, she also thinks that these skin care apps feed on the insecurity of their customers to sell their products.
It is important to note, however, that this dermatologist has a website of her own that requires customers to answer questions to find skin care solutions. But she doesn’t use this website to make money and manages it manually to offer personal advice for solving skin problems.
Another dermatologist is more positive. She thinks that these skin care apps are making people pay more attention to their skins. Her only complaint is that these apps and the solutions that they offer are too expensive. One of Proven’s kit – which includes a sunscreen lotion, cleanser and a night cream – costs $195. Skin360’s 30x zoom scanner costs $50. Few are free to use like Olay’s Skin Advisor.
It’s great if these apps can help customers with skin care solutions. But if they don’t then customers lose a lot of money in the process and delay their treatment. Perhaps, these discontented users would have found a better and quicker solution if they saw a dermatologist.
AI in Dermatology
All of this is, of course, not to discredit the contributions artificial intelligence is making to dermatology. Dermatologists do use artificial intelligence technology to detect deadly diseases like skin cancer. And these skin care apps are bound to get better with time as they gather more data and learn from their interactions with their customers. Perhaps, it’s too early to judge them then. And, you’re probably more capable of finding a solution by yourself or by seeing a dermatologist for the moment.