Anonymised, illustrative composite. A one-click AI photo tool cleaned up a kitchen ceiling for MLS — and cleaned a real water stain right out of the picture along with the clutter.
At a glance
A listing agent ran a batch of kitchen and living-room photos through an AI photo-editing tool before MLS upload — a routine step for brightening dim shots and clearing counter clutter. The tool's automatic "blemish removal" setting, applied to the whole batch without a photo-by-photo review, also smoothed out a visible water stain on the kitchen ceiling in one frame.
RECO Bulletin 5.1 sets the advertising content standard directly: claims — and by extension, the images that stand in for a claim about the property's condition — must be “factually correct, accurate, and verifiable.” CREA's own AI guidance addresses the same ground in principle: “the adoption of AI does not alleviate the professional responsibilities of REALTORS®” and REALTORS® “must remain fully accountable for the information, advice and services they provide to clients” — standards CREA states are already reflected in REALTOR® Code Article 13 (Advertising: Content and Accuracy) and Article 15 (Advertising Claims), which apply to AI-altered content the same as to any other listing photo. Nothing about running a photo through an automated tool removes the accuracy duty — it just makes it easier to miss what changed.
The mismatch surfaced two days after the listing went live, when a buyer's agent brought a client through for a private showing and the stain, plainly visible on the ceiling, didn't match a single photo in the online gallery. The buyer's agent raised it directly with the listing agent on the spot rather than filing anything formally. The listing brokerage's managing broker pulled the full photo set the same afternoon, compared it frame by frame against a fresh, unedited set, and confirmed the stain had been edited out in exactly one of fourteen photos — not a deliberate edit, but a default setting nobody had reviewed before upload.
The stain itself was a patent defect — visible to anyone who walked through the kitchen, and not the kind of thing a seller has a positive duty to disclose beyond what's plainly there to see. That is exactly why this case turns on advertising accuracy rather than a disclosure failure: the problem was never that the buyer couldn't find the stain in person — they did, in under two minutes. The problem was that the online listing, the version most prospective buyers see first and decide whether to book a showing from, no longer matched reality. That gap is squarely what Competition Act s.52 and its civil-track counterpart, s.74.01(1)(a), cover: a representation to the public that is false or misleading in a material respect, full stop, regardless of intent or whether anyone was actually fooled by it.
The listing brokerage pulled the edited photo set the same day, ordered a same-week reshoot at $275, and relisted with the accurate images. No RECO complaint was filed — the buyer's agent had raised it informally and the fix was fast enough that it never escalated. The listing agent received a documented internal compliance note, citing both Bulletin 5.1 and CREA's AI-image guidance — the same kind of synthetic-media disclosure question a doctored photo raises — and the brokerage added a rule going forward: AI photo edits get a before/after review against the original, unedited frame before any listing upload, not a batch-apply-and-forget workflow.
Had the mismatch surfaced after an accepted offer instead of during a showing — say, at a home inspection, with the buyer already committed and the stain now reading as something concealed rather than something a tool erased by accident — the conversation stops being a quiet photo swap and becomes a buyer questioning what else in the listing was cleaned up before they saw it. A $275 reshoot, caught before an offer, is a rounding error next to a deal renegotiated or walked over a credibility problem that a same-week fix would have avoided entirely.
The tell was the mismatch itself, and it was findable before the listing ever went live: a straight side-by-side of the AI-edited output against the original camera file, frame by frame, would have shown the stain missing in seconds. It only reached a buyer's agent in person because nobody had done that comparison first — the batch tool was trusted to brighten and declutter, and nobody checked what else it had quietly decided to fix. A related AI-accuracy miss, this time in text rather than an image, is an AI translation that changed the meaning.
A 30-minute call is enough to tell you whether AI pays for itself here.
