Maintenance
Before and after timelines in 2027: provenance labels, AI-image rules and US advertising law
Four shifts for cosmetic before and after photo pairs in 2027: capture smoothing, Content Credentials labels, EU AI Act Article 50 transparency rules and FTC rules on results advertising.
What to take away
- The timelines trend is the format itselfone before frame, one after frame, published as proof that something changed. In cosmetic work it is the main sales instrument, for clinics and for the celebrity pairs you scroll past.
- Before and after timelines trends 2027 come down to one problem: the before shot is no longer a neutral file. On many phones, smoothing runs at capture, before anyone chooses to edit.
- A published pair compares two processing pipelines, not two faces. Looking harder cannot recover what the camera did.
- What changes by 2027 is provenance metadata, not ethics. Cameras and platforms increasingly read signed capture records, while most front cameras write none.
- The EU AI Act's transparency rules took effect on 2 August 2026, with a grace period to 2 December 2026 only for the machine-readable marking duty of generative systems already on the market, so 2027 is the first full year in which AI-generated images in the EU carry labelling duties.
- The binding rules concern claims, not pixels. The FTC health products compliance guidance and the FTC's Endorsement Guides cover US sellers. The Ontario Medicine Act covers medical practice in that province.
- For yourself, the format still works. Fix distance, lens, light, time of day and filter, then shoot on a schedule.
Why the before shot stopped being neutral
Open your camera app and check what is switched on before you press anything. On many phones the front camera smooths and brightens at capture, and that setting can be on by default, so check yours. Look for it in the camera app's own settings, labeled beauty mode, face retouch, or skin smoothing; some apps show no indicator at all.
What reaches you has passed four stages. Two of them, capture and publication, can change a face without a person deciding to.
- Capture.The camera app or the platform applies smoothing, brightening and face shaping. Work that once belonged to a deliberate stage of image editing now finishes before the file exists.
- Selection.One frame survives out of many. In a professional shoot that is a choice among hundreds.
- Editing.Color grading and retouching. Ordinary commercial practice.
- Publication.Cropping and recompression for a feed, which shift apparent sharpness and texture again.
Four stages that change a face
- Capture: camera app smooths and shapes
- Selection: one frame chosen from hundreds
- Editing: color grading, retouching, cleanup
- Publication: cropping and compression for feed
The comparison format assumes stage one is neutral. It has not been neutral for a decade, and each camera app update widens the gap. What a pair of frames can and cannot tell you is set out in before and after timelines.
What is actually changing in 2027
The concrete 2027 change is provenance metadata, not a new ban.
Content Credentials, the standard maintained by the Coalition for Content Provenance and Authenticity, attach a signed record of capture and edits to a file. Leica, Sony and Canon, among others, have shipped cameras that write it. Adobe and Microsoft read it.
Meta began labeling images it detects as AI generated in 2024, using an "AI info" label. TikTok applies automatic labels to posts arriving with Content Credentials attached. The Content Credentials mark itself is a small "cr" icon that opens the file's capture and edit history.
In the European Union, the main dates are now set. Article 50 of the AI Act (Regulation (EU) 2024/1689) sets the transparency duties, and they apply from 2 August 2026. Under the Digital Omnibus on AI, which was adopted in mid-2026 (the Commission's factsheet gives the grace period as until December 2026), generative systems already on the market before that date have until 2 December 2026 to add machine-readable marking to their outputs. Systems placed on the market on or after 2 August 2026 get no grace period, and the labelling duty for deep fakes is not deferred. Under those rules, deep fakes should be clearly labelled as artificially generated or manipulated. Supervision sits mainly with national authorities in each member state, and the Commission is preparing a code of practice on marking and labelling AI-generated content. Check the Official Journal text before relying on any date. For a cosmetic pair, that means an AI-generated "after" image published in the EU in 2027 is expected to say so; a retouched photograph is a separate question.
None of that repairs a phone selfie. Front cameras on the handsets most people own still write no provenance, so the amateur pair stays unverifiable while professional files become checkable.
Expect the labels to be uneven. A clinic shooting on a signed camera can show its before frame is unprocessed. An account posting two phone frames cannot, and a missing label proves nothing either way.
Professional regulators push from the other side. Why a qualified practitioner should decline to read a stranger's face is in beauty expert analysis.
Which documents govern a published pair
Three documents matter more than any trend piece. The FTC's health products compliance guidance sets out what a seller must prove about a health claim. The FTC's Endorsement Guides add a rule aimed at the format itself: their example says an ad that simply features before and after pictures of a person reporting a result is likely to convey that the result is representative of what consumers will generally achieve. Ontario's Medicine Act governs medical practice in that province only, and the US rules are set out in full in before and after photo advertising rules; and the College of Physicians and Surgeons of Ontario is the regulator that holds physicians to it. It matters for an Ontario clinic's photos, not for a US one.
For a US cosmetic before and after advertisement, the base law is Section 5 of the FTC Act, which prohibits deceptive advertising, and the Endorsement Guides in 16 CFR Part 255, which govern how testimonials and typical results may be shown. State medical boards add their own advertising rules for licensed practitioners. A claim that a pair shows a typical result needs evidence that the result is typical.
| What is at issue | Document | Who applies it |
|---|---|---|
| A health claim attached to a treatment | FTC health products compliance guidance | US Federal Trade Commission |
| A physician's practice and public conduct in Ontario | Ontario Medicine Act | College of Physicians and Surgeons of Ontario |
| Whether a file was altered after capture | Content Credentials, C2PA standard | Camera makers, editors, platforms |
A regulator can punish a false claim. No regulator can tell you whether the pixels were smoothed.
Neither route reaches a private account posting two selfies.
What a published pair can and cannot support
A pair drawn from two pipelines carries an unknown at its center. You cannot inspect a published photograph and recover the camera's processing. The confident caption underneath is unsupported.
Clinic galleries are the case worth pressing, because you can ask direct questions. Same camera, same lens, same lighting, same interval, any retouching? Vague answers mean the images are decoration.
The errors that catch experienced operators, even in good galleries, are covered in seven things to do instead of a photo pair.
Doing it properly on yourself
The format works on your own change if you control every variable. Pick something you will follow for months: a routine, a treatment, or healing after a procedure.
- Fixed distancemark the floor and stand on the mark.
- Same lensthe back camera, not the front one.
- Same lightone lamp, same position, no daylight in frame.
- Same time of day, within about an hour.
- Smoothing, beauty mode and filters off, checked in the app settings.
- No edits after capture. No crop, no color, no retouch.
Controlled self-tracking setup
- One camera, one spot, one distance
- Same daylight, same time, never overhead
- Beauty modes and filters turned off
- Camera at eye level, same expression
- Same makeup or none in every frame
- Keep the whole series, not best and worst
- Note date and what changed, one at a time
Shoot on the same interval until the change stops. Compare the newest frame against every earlier one, not only the first. A slow series is less dramatic than the ones you scroll past. It is also the only version that produces information.
If a comparison sent you looking at treatment, cost and consent questions matter more than the pictures. The reasoning in hair transplant mistakes applies to any procedure sold with photographs.
