Can you trust AI skin analysers?

We've spotted AI skin analysis tools popping up everywhere this year, from big retailers such as Boots and Superdrug to individual beauty brands, so we wanted to put them to the test.

What we found when we tried one out was that results aren't as intelligent as you might hope, and while some recommendations may be helpful, there are pitfalls to watch out for - and some suggested products may not be appropriate for your skin.

We also found some brands came up far more frequently than others in the list of recommended products the tool churned out.

How do AI-powered skin analysers work?

AI skin scanners claim to offer personalised skincare insights along with product recommendations based on a quick selfie that you upload to the tool. In theory, they use algorithms and machine learning to take a look at your selfies and scan for things like wrinkles, breakouts, and dark spots. 

You may have spotted them on online pharmacy sites and beauty store apps, promising instant feedback and a list of recommended products tailored to your concerns and skin type. But while they seem convenient,  we found the results can be inconsistent and prone to providing inaccurate recommendations.

On top of that, differing skin tones and poor lighting can easily mess with the results, raising false alarms or nudging you to buy skincare you don't really need.

Health? or Hype?

Make the right choice with recommendations you can trust, backed by expert testing.

Explore health recommendations

Our AI skin scanner test

We identified 17 online AI skin scanning tools on various sites such as Boots and Superdrug, as well as major skincare brand homepages like La Roche-Posay, L'Oréal and Neutrogena. 

Many of these services share the same or similar underlying technology: at least seven popular options run on a white-label AI engine created by Revive (including Boots). Some major skincare brands have their own version, such as L’Oréal's proprietary 'ModiFace' framework. 

We asked 30 people to test out the Boots Skincare Diagnostic Tool, running over 120 individual scans and tracking hundreds of data points. This is the most mainstream option, run by the most common AI skin analysis platform.*

• Everyone filled out the pre-scan questionnaire, carefully noting down their actual skin concerns, skin type (e.g dry or oily), age group, gender, and location.
• Next, they did two back-to-back scans and logged what came up each time.
• These first two scans were done right after each other in great lighting, keeping conditions as identical as possible.
• Then came a third scan, but this time our testers intentionally gave inaccurate answers on the pre-scan questionnaire to see if it changed anything.
• Finally, everyone ran one last scan in dim or patchy lighting to see how this affected results.

After each scan, participants noted the app's overall skin score, its analysis breakdown, and any product recommendations that appeared.

*Note that since we tested it in early September 2026 and informed Boots of our findings, the Skincare Diagnostic Tool has been removed from the website.

Get expert tips on the skincare products you do and don't need in our podcast episode

What our AI skin scanner testing found

When lighting and distance were kept identical, the tool gave consistent results. Each scan provides the user with an overall numerical score and this score barely shifted between the first and second scanes — by less than a single point on average. Only four out of 30 testers saw their score change by 5 points or more. What's more, their top and bottom scoring categories (e.g scores for redness etc) and suggested products, stayed almost the same.

It struggled to catch self-reported localised skin issues like acne, breakouts, redness, and hyperpigmentation, though. Out of 14 people with visible acne and breakouts who explicitly listed them as their primary concern, the scanner failed to spot this or directly contradicted them 10 times. 

Similarly, for nine testers with dark spots, the app rated their pigmentation as perfectly healthy six times. As for redness, out of seven affected testers, not a single one had redness flagged as a concern, and three were even awarded top health scores for clear skin.

All in all, about 63% of testers received scores that matched what they saw in the mirror, while 37% got results that completely mismatched their actual skin concerns.

Good lighting makes a big difference

In fairness, the app warns you that good lighting is key for accurate results — it won't even snap a photo if the room is too dark or the angle is awkward. 

Our tests confirmed just how much difference it makes: for about half our group, simply shifting light sources caused overall skin scores to swing wildly, jumping up to +39 points or dropping by as much as -25 points.

Even subtle changes—like light coming from the left instead of the right—noticeably affected the final results. 

Skin tone can influence the accuracy of the results

Skin tone plays a huge role in how accurate these scanners are. Many AI-based tools struggle to accurately diagnose people with darker skin tones (Fitzpatrick Types IV-VI) because datasets often underrepresent them. 

In our tryout, the app struggled to detect hyperpigmentation and dark spots on participants with darker skin tones (Fitzpatrick Types IV and V), giving them higher overall health scores on average than those with lighter skin types.

Lighter skin tones experienced bigger score swings when moving between good and bad lighting.

AI skin scanner product recommendations aren't always appropriate

Before taking a photo, you answer a quick questionnaire about your age, gender, skin type, location, and key concerns like fine lines or breakouts.

We found that your answers don't change your overall score much, but they massively skew category-specific ratings. For example, one tester with clear skin, which the scanner correctly rated as having little to no acne, manually selected spots as a main concern in the third scan; the scanner then gave them a low acne score, indicating they did, in fact, have acne. 

While this calls into question the connection between the scanner and the product recommendations, trusting what you type over what the camera sees is theoretically safer, since it could, in theory, prevent the app from recommending harsh ingredients that could irritate sensitive skin. Unfortunately, however, this didn't stop the tool from recommending products that could aggravate existing conditions. In one case, a tester with rosacea was recommended products containing retinol and salicylic acid, which could irritate their skin. 

You might think it's the AI wizardry determining your recommended products, but we found that self-reported quiz answers heavily shape product picks: 25 out of 30 testers got completely different morning product lists just by tweaking their pre-scan answers.

Some brands seem to be recommended more frequently than others

By default, the Boots AI tool suggested four products for morning use (cleanser, serum, moisturiser, sunscreen) and five for the evening (cleanser, serum, night cream, eye cream, mask). You can filter results by things such as vegan, paraben- or fragrance-free options, or budget, though our testers logged the default recommendations.

A couple of brands dominated the recommendations given to our 30 testers. L'Oréal and Cosrx were suggested far more frequently than other brands —214 and 154 times respectively — while other popular brands like Olay and La Roche-Posay barely showed up.

We got in touch with Boots to ask how these product recommendations are chosen; they didn't answer our specific questions but did inform us that the Skin Diagnostic Tool was no longer available online. 

Expert view on free AI skin scanners

We reached out to consultant dermatologist Dr Emma Wedgeworth to get her take on what you should keep in mind when using these tools.

Best anti-ageing creams:  We reveal which face creams really made a difference to wrinkles and more

What some of our testers thought

We asked our testers about their experience of using the Boots Skincare Diagnostic tool. Some were impressed by the fact that the app recommended products they already used:

'I was quite impressed by the product recommendations as I use a vitamin C serum during the day. It also recommended the sunscreen I actually use, which gave it extra credibility.'

When it came to the accuracy of the scans though, the response was mixed. 

'The second scan didn't register any spots even though I could see them on my face; the app also didn't seem to register any hyperpigmentation even though I know I have it.'

'Overall, the results were broadly accurate in identifying my main concern of dark spots and hyperpigmentation, which I genuinely experience. My skin was otherwise assessed relatively positively. However, some of the other results and product recommendations varied between scans, suggesting the tool may be less consistent when assessing individual skin concerns and recommending products.'

We did see one instance of a participant being recommended products that could aggravate their skin conditions: 

‘I'm amazed. I have rosacea and am recovering from a bit of an outbreak, so I have some spots. I said spots and blemishes were my main concern and that I had sensitive, dry skin. Yet I have scored highly for spots and been recommended products that contain things like salicylic acid and retinol, which I think would make my skin so much more irritated. I guess this tool is not for people with skin conditions, as I have obviously visible rosacea!’

Should you use AI skin analysers?

We noticed that several high profile skin diagnostic tools, including the Boots one, have recently been pulled from the retailer websites. But for those that remain, we think it's worth exercising caution, especially if you have existing skin concerns.

Calling these tools 'diagnostic' sets unrealistic expectations for anyone looking for actionable advice on skin issues such as sensitive skin, acne, redness, or hyperpigmentation, and it's important to realise their limitations.

They can be fun to try out and help narrow down your choices, but be aware it might not be clear why you are being offered specific products, and when it comes to actual skin health concerns, it's no replacement for seeing a dermatologist or medical professional. 

If you do use one, make sure you fill out the questionaire carefully, take your photo in good lighting, and do additional research on any recommendations you aren't sure of.

Do LED face masks really work? Get our expert assessment and find out which masks have impressed in our tests so far