ChronoVerify Guides
Photo verification, in plain language
Practical guides to checking when a photo was taken, what its metadata can and cannot prove, and how to tell signed or AI-generated images apart. Written by a veteran intelligence analyst.
What AI-presented images actually carry: 73 images, 38 companies, August 2026
We fetched 73 images that 38 companies publicly present as AI-generated or AI-altered, exactly as each file is delivered, and checked them for Content Credentials against the official trust list. 72 carried none. One arrived with a validated credential, and the delivery path is the reason.
Read the dataChronoVerify joins Google, Microsoft, Meta, and OpenAI as initial signatories of the EU Code of Practice on Transparency of AI-Generated Content
The European AI Office published the first signatory list on July 31, 2026: about 190 organizations, 83 under Section 1. ChronoVerify signed as a detection-side technology provider, two days before Article 50 applied. What that covers and what it does not.
Read the updateChronoVerify is a C2PA Conformant Validator: what the listing covers, and what it does not
ChronoVerify is on the C2PA Conforming Products List as a conformant validator product, record 019f8a20-6452-7a43-b11b-59d0b0e4a84a. What the record says, what conformance is not, and why the list's 124-to-20 split between writing and reading is the story.
Read the updateCalifornia AB 853: the law that makes platforms read provenance, not just write it
Almost every AI transparency rule asks whoever generates content to mark it. AB 853 is the exception: from January 1, 2027 a large online platform must detect the provenance data in what it distributes, and say whether the file came out of a model or out of a camera.
Read the guideEU AI Act Article 50: a practical checklist for your image pipeline
Who the transparency regime affects, what the provider and deployer duties require, and a step-by-step checklist to stand up an auditable provenance step this week, from one keyless call to signed, timestamped audit records you keep in your review files.
Read the guideChronoVerify update, July 18, 2026: an Article 50 marking checker and richer reports
A free Article 50 marking checker joined the tools fifteen days before the August 2 deadline, signed reports now show the analyzed image with a plain-language verdict scale, and the verifier previews your photo without uploading it anywhere.
Read the updateChronoVerify update, July 17, 2026: trusted timestamps in every signed report
Every signed report now embeds an RFC 3161 trusted timestamp from an independent authority, validated locally before it is claimed and verifiable offline with openssl. Plus a quickstart hub at /docs and a new brand mark.
Read the updateChronoVerify update, July 2026: what shipped and who it is for
Verdict permalinks, a sample gallery with known ground truth, raw GPS coordinates for OSINT work, two iPhone photo bugs found by real users and fixed within hours, and a plain list of who image provenance verification helps.
Read the updateChronoVerify joins the Content Authenticity Initiative
We are now a member of the CAI, the community behind C2PA Content Credentials. What membership signals, the boundaries we hold, and how it fits our cryptographic validation.
Read the updateContent provenance for AI systems: why validation beats detection
As machines make and read more images, verifiable provenance, not probabilistic detection, is the durable trust signal. What AI pipelines need, and how Article 50 makes it the baseline.
Read the guideHow to tell when a photo was taken
Where the capture date lives, how to read it on any device, and how far to trust it.
Read the guideWhat are Content Credentials (C2PA), and how do you check them?
A signed C2PA record of how an image was made and edited: what it holds, who signs it, and how to verify one.
Read the guideHow to read EXIF metadata, and what it can and cannot prove
Read a photo's date, device, and GPS on any device, and why EXIF is a lead, not proof.
Read the guideIs this photo AI-generated? How to actually check
Pixel AI-detectors misread real photos as fake. Check provenance (C2PA and EXIF) instead.
Read the guide