Detect AI-generated and edited claim photos

Yes. Trustable checks each claim photo for traces that image generators leave, local edits such as added or enlarged damage, and gaps in the camera data and file history. Each finding is marked on the photo with its strength, so an adjuster or special investigations unit (SIU) can see why a photo was flagged before the claim is paid.

A shallowfake is a real photo or document changed with an ordinary editing app rather than generated from scratch by AI. Common examples are damage painted onto a car photo, a date changed on a receipt, or an amount raised on a repair invoice. Because most of the file is genuine, shallowfakes can pass a quick look; pixel-level and document checks are designed to flag them.

Yes. Trustable checks repair invoices and receipts for edited text, font and layout mismatches, rebuilt templates and line items or totals that don’t add up. Each finding is marked so the adjuster can follow up with the repair shop. Comparing an invoice with the rest of the claim, such as its dates and the damage in the photos, is in development.

Through the API or the Trustable web app. Send damage photos, invoices or receipts and get back the findings, where they are and how strong each one is, for your adjuster or SIU. Routing results into your claims system is in development.

No. Messaging apps, screenshots and resizing often strip camera data from genuine photos, so missing metadata on its own proves nothing. Trustable treats it as one signal among several, weighs it with pixel-level and AI-generation checks, and shows how strong the signal is, so the adjuster isn’t pushed to deny a valid claim.

Yes. Recorded claim videos and video-call recordings are checked for face swaps and other signs of generation or editing, and each finding is marked on the frames where it appears with its strength. As with photos, a finding is a reason for the adjuster or SIU to ask for more evidence, not a decision on its own.

No. Content Credentials (the C2PA standard) show that a file’s recorded history is intact, such as which device or app created it. They don’t prove that the loss in the photo happened, and many claim photos carry no credentials at all. Trustable checks the photo itself either way; reading credentials is in development.