Editor's note: Please seek advice on your individual skin health from a licensed health care provider.
Over half of Americans use free versions of AI models like ChatGPT on a daily basis. However, last year, nearly half of its responses to questions about common skin conditions were deemed inaccurate or too vague to be useful in clinical practice.
From “How do I stop my skin breaking out?” to “What can I do to reduce redness?”, even seemingly straightforward skincare questions can produce unreliable answers. If AI currently struggles with basic dermatology questions, why are we trusting it to tell us what our skin needs?
A quick, confident answer isn’t the same as a correct one
We know why people
are turning to AI. It's instant, it's often free, and getting in to see a dermatologist takes an average of 36.5 days and costs around $245. Upload a selfie, describe a breakout and within seconds you can have a list of possible causes and products to try. In 2026, 34% of U.S. adults said they had used an AI chatbot for at least one health reason, and speed and cost were among the reasons they gave.
But convenience isn’t the same as credibility. AI can help us understand ingredients, compare products and access information, but there’s a difference between using it to become better informed and trusting it to make decisions for us. A quick, confident answer isn’t necessarily an accurate one, and allowing AI to dictate what our skin needs can lead us to act on advice that isn’t right for us.
AI-powered skin care tools could help people better understand their skin and find products that suit their needs, but not all tools have equal merit. One in five Americans are turning to AI chatbots for health advice, despite general-purpose large language models (LLMs) not being designed to assess individual skin concerns. The challenge is knowing when AI can be a useful starting point and when its recommendations need to be questioned.
LLMs are designed to generate a plausible, decisive response from any information given to them, not to assess your symptoms, medical history and lifestyle like a skin care expert would. They draw on everything from trustworthy scientific research to unfounded online discussions and commercial content, without always making clear what's reliable and what isn't.
LLMs also frequently accept the assumptions built into the questions we ask them. Ask what’s causing your “hormonal acne” and the answer may start from the assumption that it is hormonal acne in the first place. Adding to this lack of reliability, often the quality of a response is only as good as the quality of the prompt given to it. It’s easy to mistake a confident chatbot answer for an expert one.
There’s another factor to consider: Skin doesn't exist in isolation. A photo can be a useful starting point for understanding your skin, but it only tells part of the story. Sleep, stress, diet, hormones, medication, and lifestyle can all influence how it looks and behaves. Two people might have almost identical-looking breakouts for completely different reasons. Without context, a photograph alone cannot tell us whether someone is dealing with an underlying condition, a hormonal change or a reaction to a product.
Children and teenagers are especially vulnerable. Many are experimenting with skincare, sometimes without their parents’ knowledge, and using potent active ingredients, including retinoids and exfoliating acids, without necessarily understanding whether they're suitable for their skin or how to use them safely.
AI may recommend products aimed at adults without accounting for age differences or the risks of layering multiple actives. This makes ingredient awareness and age-appropriate skincare essential, as the wrong products can cause irritation, sensitivity and even worsen breakouts, considerations AI advice often fails to acknowledge.
It can give you a convincing answer, but it doesn’t necessarily have the reliable information, personal context or clinical judgement needed to tell you what your skin actually needs.
Are we personalizing skin care or just finding more flaws?
AI chatbots like ChatGPT risk teaching people that ordinary features are now problems to fix. Pores, pigmentation, redness, fine lines and uneven texture are identified as concerns, encouraging users to constantly do more to their skin.
Adding another exfoliant, another active, another serum to target the latest concern can compromise the skin barrier, leading to over-exfoliation, redness and irritation. Some dermatologists have warned that AI chatbots are exacerbating this issue, with reports of increased irritant contact dermatitis and rosacea flare-ups after patients layered multiple chatbot-recommended products.
My greatest concern is that this newfound trust in unregulated AI-powered skin care advice is increasing damage to skin, not reducing it.
Our skin does not need to meet an algorithm's definition of perfection. Before asking AI what is wrong with your skin, take a step back and ask whether there is a problem at all. Understand what you're putting on your skin, introduce products for a reason and be wary of recommendations that encourage you to keep adding more. AI can be a useful starting point, but it shouldn't replace your own judgement about what belongs in your routine. Sometimes, the best thing you can do for your skin is keep things simple. Your skin needs less than you think.
Daniel Hoftun is the founder of all-natural skin care brand Norse Organics.
This article originally appeared on USA TODAY: Tempted to ask AI about your skin? Think again
