Your AI says you need fillers. Your AI has never touched a face

Chatbots are now writing treatment plans for Dubai's aesthetic patients. A medical director explains why that should worry all of us — and where AI really belongs.

Every week, a patient sits across from me, unlocks her phone and shows me a screenshot. “The AI says I need 2ml in my lips, Botox in my forehead and a thread lift.” She arrives with a plan. What she doesn't have is a diagnosis.

I welcome informed patients. Curiosity is healthy. But aesthetic medicine has entered a strange new era: the chatbot consultation, where a machine that has never seen your skin under clinical light, never felt the tissue and never asked about your medications hands you a treatment plan with total confidence.

Speaking at length about AI and Chatbots, Dr. Maya Khoury, Co-Founder & Medical Director, Le Rêve Clinic, Dubai said, “Here is the uncomfortable truth: AI does not tell you what you need. It tells you what your prompt asked for. Ask “what's wrong with my face?” and it will find something wrong. Upload a filtered selfie taken at arm's length under poor lighting, and it analyses the filter, the lens distortion and the shadows — not you. Phrase the question with a hint of insecurity, and it mirrors that insecurity back. These tools are built to be helpful and agreeable. In medicine, agreeable is not the same as safe.”

What a chatbot cannot see

A photograph is two-dimensional. A face is not. Bone structure, muscle movement, skin thickness, vascular anatomy — the things that decide whether a treatment looks beautiful or goes wrong — sit beneath the surface. The AI doesn't know you had filler three years ago that never fully dissolved. It doesn't know about your medical history or the medication you take. It doesn't know your asymmetry is skeletal, not soft tissue. And it doesn't know that the “flaw” it found is what makes your face yours.

Mistakes have consequences: overfilled lips, frozen expressions, migrated filler and, in rare cases, serious vascular complications. A chatbot carries no responsibility for any of them. Your doctor does.

AI belongs in the clinic — in the right hands

So should we reject AI? Absolutely not. The problem is not the technology; it is who feeds it, with what, and who interprets the answer. AI is only as good as the data it receives. Standardised imaging, calibrated lighting, a full medical history and a trained clinician asking the right questions — that is what turns AI from a mirror of our insecurities into a genuine diagnostic tool.

That is why at Le Rêve, which I helped found, we are introducing full AI facial and skin scanning built on approved medical devices qualified for clinical assessment. These systems measure what the eye can miss — texture, pigmentation, volume, symmetry and change over time — under controlled conditions. But the machine does not make the decision. The doctor does, with the patient, in a real conversation.

A message to patients

Bring your questions. Bring your screenshots, even. But please don't bring a prescription written by an algorithm and ask me to execute it. You would never ask a surgeon to operate from a Google search. Your face deserves the same respect.

The best aesthetic result has never been the one a machine recommends. It is the one where you still look like yourself — only rested and confident.