Photo Proof for Pros: Defend Your Work Against AI

Lawyers, insurance adjusters, real-estate agents, tradespeople: a photo was long enough to back up a case. That’s no longer true. Here’s why, what the law actually says, and how to give your images evidentiary value again.

"That photo — who’s to say it wasn’t doctored?"

You photograph water damage to file an insurance claim. Three weeks later, the adjuster replies: how can we be sure these shots really date from the incident, and that they haven’t been retouched? You have no answer. Your photos are genuine — but you can’t prove it.

This scene now repeats itself in every profession where images serve as evidence. A lawyer producing a visual exhibit, an agent documenting a property inspection, a tradesperson justifying a before/after: the opposing party’s first reaction is no longer to look at the image, but to cast doubt on it. And that doubt alone is enough to weaken a case.

Why a photo no longer proves anything

The distrust isn’t irrational. Deepfake fraud rose by roughly 700% in a single year between early 2024 and early 2025, with global losses estimated at nearly €863 million in 2025 alone. In 2024, a deepfake attack was recorded every five minutes. With consumer tools like Sora or Veo, faking an image or a video no longer requires any technical skill.

The real problem is more insidious than fakes circulating. It has become as easy to fake an image as to claim that a genuine image was faked. In other words: the fake contaminates the real. Your honest photos lose their force the moment no one can rule out the possibility that they were fabricated.

Does a photo have legal value? What the law really says

In civil matters, evidence is unrestricted: it can be brought by any means. A photograph is therefore perfectly admissible. But admissible doesn’t mean decisive: the judge has sole discretion over its evidentiary weight. A photo isn’t absolute proof — it’s an element of proof whose weight depends on the trust placed in its date, origin and integrity.

Yet an ordinary digital photo is fragile on all three points: no guarantee of integrity, and metadata editable in a few clicks. That weak link is exactly what the opposing party exploits to dismiss your exhibit.

"We’ll get a bailiff’s report if needed" — except you can’t turn back time

The seasoned professional’s instinct is the bailiff’s report (constat de commissaire de justice). And rightly so: it remains the strongest evidence to produce before a court or an insurer. But it has two practical limits. It’s too slow and too costly to trigger in every situation. And above all, it can’t be done retroactively: the day you observe the damage, the inspection or the work, that’s the day you need proof — not three weeks later when the dispute erupts.

Hence the need for an intermediate level: a timestamped, verifiable record, created at the moment of capture, that strengthens your photos without mobilizing a public officer for every shot.

Detecting fakes won’t solve the problem

You might think it’s enough to analyze suspicious images with an AI detector. But these tools are unreliable: MIT calls them "neither accurate nor reliable", their false-positive rates often exceed 10%, and OpenAI shut down its own detector for lack of accuracy. A detector doesn’t identify an origin; it estimates a probability — a race lost in advance against generators that improve every month.

The only robust approach reverses the logic: rather than guessing after the fact whether an image is fake, you certify its authenticity at the source.

Certifying at capture: the approach already adopted by Adobe, the BBC and Microsoft

This isn’t a startup whim. Since 2021, the C2PA coalition — Adobe, Arm, BBC, Intel, Microsoft, Truepic and over 200 members — has been developing an open standard, Content Credentials: a cryptographic fingerprint (a hash) of the content, signed and attached to the file, that any modification invalidates. The standard is in the process of being adopted as an ISO norm. Tomorrow’s proof of authenticity isn’t sought after the fact: it’s laid down at the moment the image is born.

Truth-Check: certify a photo or video in three steps

Truth-Check applies this logic in a simple app, designed for freelancers and small businesses — not just large groups. The principle:

You capture the photo or video from the app.
Truth-Check computes a cryptographic fingerprint of the original content and links it to the date, time and device metadata (geolocation optional).
You get a timestamped certificate and a public verification link, shareable with a client, a colleague, an insurer or a court.

An important point of honesty: like the C2PA standard, Truth-Check certifies the origin and integrity of content — not the "truth" of the scene. It proves that an image was captured at a given moment and hasn’t been modified since. That’s precisely what an ordinary photo lacks, and what makes the difference in a case. The app doesn’t replace a bailiff’s report: it works upstream and as a complement, where a formal report would be disproportionate.

What to do, concretely, from your next case onward

Next time you document an incident, an inspection, a job or an exhibit destined for a dispute, don’t just take a photo: certify it the moment you take it. You turn a contestable snapshot into a timestamped, verifiable record — and you cut short, in advance, the "nothing proves this is genuine".

Go further: read our guide on photos as evidence for lawyers and real estate, our guide to detecting AI-generated photos, or try our free photo verification tool.