In short. Where the line falls between describing an image and naming a disease, and how that line is enforced across different markets.
Key points
- A measurement shown to a patient and a diagnosis given to a patient are separated by a regulatory boundary, not by a matter of degree.
- A product that displays a measurement while its marketing implies a diagnosis has not been assessed for the claim its marketing makes, and that gap is the single most useful thing to look for.
The regulatory position of software in aesthetic medicine turns on a single distinction: whether the product describes an image or makes a clinical claim about a person. Most of the confusion in this area comes from products that do the first while being marketed as though they did the second.
Describing an image
A tool that segments an image, measures a distance, tracks a change between two photographs or highlights a region is making a technical claim. It is regulated as a device for that named task, and the evidence it must produce is about whether it performs the task reliably. That is a genuine requirement and it is much smaller than the requirement for a diagnostic claim, because the output is a property of an image rather than an assertion about a person.
Naming a condition
A tool that outputs a diagnosis is making a clinical claim. It must show accuracy in the population where it will be used, which means sensitivity and specificity measured on real patients rather than agreement with a panel on a set of photographs. This is a considerably higher bar, and it is the bar that a photograph-based consumer app has almost never met.
Describing is not diagnosing
A measurement shown to a patient and a diagnosis given to a patient are separated by a regulatory boundary, not by a matter of degree. A product that displays a measurement while its marketing implies a diagnosis has not been assessed for the claim its marketing makes, and that gap is the single most useful thing to look for.
Claiming nothing at all
Some products fall outside device regulation because they claim nothing regulated. A rendered prediction of a surgical result is generally in this group: it illustrates rather than diagnoses or treats. An app that returns generic advice is also in it. Falling outside regulation is not evidence of harm and it is also not evidence of anything, and it should not be read as a clearance that was granted.
Why jurisdictions diverge
What counts as a medical device, what counts as a claim, and how much evidence a class requires all differ between markets. A tool cleared for measurement in one jurisdiction may need a separate route in another, may be classified differently, or may not be marketed at all. Marketing material that cites an approval without naming the market and the task is citing something that may not apply where you are reading it.
What to ask about a regulatory statement
- Which authority granted it, and for which named task or indication?
- Does the claim being made to me stay inside that task, or has it been extended in the telling?
- Is the product regulated at all, or is it outside device regulation because it claims nothing?
The boundary is administrative rather than clinical, and it is the most reliable guide to what a product has actually been asked to demonstrate.
Frequently asked questions
Is AI software in aesthetics regulated?
Where it makes a medical claim, yes. Software that measures or analyses an image is regulated as a device for a named task, and software that names a disease is regulated for a diagnostic claim at a considerably higher bar. Software that claims nothing regulated generally falls outside device regulation altogether.
What is the difference between describing and diagnosing?
Describing reports a property of the image: a distance, a colour, a change between two photographs. Diagnosing asserts that a person has a condition. The second is a clinical claim, and it requires evidence of accuracy in the population where the tool will be used, which is a much larger undertaking.
Does the same product have the same status everywhere?
No. Device classes, the evidence required for each, and what counts as a claim all differ between jurisdictions, so the same software can sit in different classes in different markets or have no authorisation at all in one of them.
Risks & Limitations
- Software measures an image; it does not examine tissue, and an examination is what a diagnosis usually turns on.
- A rating reflects what a model was trained on rather than a standard of beauty, and it is least reliable for faces least represented in its training data.
- Agreement between software and clinicians is measured against an opinion, not against correctness.
- A predicted image is a drawing of a plan expressed in a body the model does not describe.
- No algorithm takes a history or carries responsibility for the advice it gives.
When to Seek Care
- A new or changing lesion, a mole changing in size, shape or colour, or a sore that is not healing — have it examined by a clinician rather than assessed from a photograph.
- A skin change you are worried about, whether or not a tool has told you it is nothing.
- A reaction, infection or worsening after treatment — contact the service that holds your records.
- If you have uploaded a facial image and want to know how it is held or want it deleted, ask the service that holds it under the terms you accepted.
References & Further Reading
- U.S. Food and Drug Administration (FDA) — guidance and authorisations for software as a medical device and for image analysis
- European Commission and notified bodies — Medical Device Regulation and artificial intelligence guidance
- National Medical Products Administration (NMPA, China) — approved software and its named claims
- World Health Organization (WHO) — guidance on the ethics and governance of artificial intelligence for health
- Cochrane Library — systematic reviews of diagnostic and decision-support software
This page is patient education, not medical advice, and it does not cite individual studies. It reviews evidence and regulation; it is not a booking page and does not recommend a treatment. Discuss your own case with a qualified clinician. To ask a question about this review, see Contact & Consultation.