Newsrooms verify images for a living, and they have done so since before the diffusion-era panic. The institutional practice that wire services, broadcasters, and investigative outlets have built over the past decade is the working model for how a verification operation can be structured. This page describes what the major operations actually do — their desks, their tooling, their decision criteria — and where their standards diverge. The intended audience is anyone designing or evaluating a newsroom-grade verification operation; for individual practitioners, the workflow page covers the methodological backbone.
The newsroom verification operation is a specific kind of editorial function: a team or desk whose job is to assess images before they enter the publication chain and to investigate questionable images that have already circulated. The function is institutional, with its own staff, budget, training, and accountability. The largest operations — at AP, Reuters, AFP, BBC, the New York Times — are full departments. Smaller outlets have part-time or distributed functions. Project Origin and other inter-organizational projects coordinate practice across them.
The major operations
Associated Press
AP's verification function is integrated into its global photo desk. Wire images are verified before being put on the AP feed, with verification including source authentication, on-the-scene corroboration through staff stringers, and where available C2PA validation. AP was an early participant in Project Origin and ran high-profile C2PA pilots in 2024 and 2025 for selected story types. AP's verification standards are the de facto baseline for wire imagery in much of the world; an image that passes the AP feed is treated as having cleared a threshold of editorial scrutiny.
Reuters
Reuters operates a similar structure, with the additional element of its 2006-era post-Hajj editorial procedures that hardened the wire-service consensus against compositing and content removal. Reuters has been particularly aggressive about C2PA integration, with credentialed wire delivery on selected story types and active participation in the C2PA coalition. The Reuters Handbook of Journalism contains specific provisions on image manipulation that have informed the broader wire-service standard.
AFP
Agence France-Presse runs a fact-check operation (AFP Fact Check) that operates partially independently of the wire desk and handles high-volume verification of viral imagery and claims. AFP has been a notable user of automated detection tools at the front of its workflow, with human verifiers reviewing flagged content. AFP's published verification methodology guides have been influential for smaller outlets.
BBC Verify
The BBC's open verification desk, launched in May 2023. BBC Verify is unusual in that it operates with substantial editorial visibility — its analyses are themselves published, with named verifiers walking through their reasoning on-air and in articles. The transparency is intentional, both as a journalistic standard and as a counter-disinformation strategy. BBC Verify uses standard tooling (InVID, reverse search, metadata, C2PA validation) within a documented editorial framework.
Bellingcat
The open-source investigations organization rather than a traditional newsroom, but its verification methodologies have shaped newsroom practice substantially. Bellingcat's published guides on geolocation, image verification, and OSINT techniques are widely used as training material. The organization's investigations of Russian military operations, MH17, and various human-rights cases have established detailed methodologies for verifying user-generated content from conflict zones, which the conventional newsrooms have absorbed.
The Washington Post Visual Forensics
A team within the Post that specializes in detailed analysis of visual evidence, often producing multi-day reconstructions of specific incidents from social-media imagery. The unit's work on the January 6 2021 Capitol attack and on conflicts in Ukraine and Gaza has demonstrated what extended forensic-journalistic analysis can produce when given time and resources. The methodology is closer to documentary investigation than wire-desk verification, but the skills overlap.
The shared workflow
Despite institutional variation, the procedural core of newsroom verification is reasonably consistent across operations. The shared steps:
- Intake. The image arrives via wire, social pickup, audience submission, or stringer contact. Source and provenance metadata are captured at intake.
- Source check. The submitter's identity is established and their previous submissions reviewed.
- Provenance inspection. Any C2PA credentials are validated and the chain is recorded.
- Metadata read. EXIF, XMP, and IPTC are inspected for consistency with the submitter's account.
- Reverse search. The image is run through multiple engines to check for prior appearance.
- Forensic check. Pixel-level forensic indicators are inspected; AI detection is run as a triage signal.
- Independent corroboration. Other sources are sought who can confirm the depicted event.
- Editorial decision. A named editor, with documented criteria, decides whether to publish, hold, or kill the image.
- Record. The verification work is archived for later defense and for institutional learning.
This is essentially the verification workflow from the workflow page with two newsroom-specific additions: source-check (steps 2) and independent corroboration (step 7). Both are journalistic rather than technical, and both are typically what distinguishes a professional verification from an automated one. The technical tooling is necessary but not sufficient; the journalistic relationships and source assessment are the part that the institutions invest in most heavily.
How decisions get made
The output of the verification process is not "the image is real" or "the image is fake." It is a verification opinion, qualified by the strength of the supporting evidence, that an editor uses to make a publication decision. The opinion typically reads something like: "Image submitted by [name], a stringer with three years of work for AP. C2PA chain valid, signed by photographer's certificate. EXIF consistent with submitter's claimed equipment and location. No earlier appearances found. Forensic indicators consistent with claimed capture. Independent corroboration from two regional staff. Recommend publication."
Or, more often, the opinion is mixed: "Submitter unknown, image found on Telegram channel. No credentials. EXIF stripped. Reverse search returns no matches. Visual content matches geolocated landmarks. Recommend publication with prominent attribution to channel and explicit caveat about source unverified." The publication decision is then an editorial judgment about whether the qualified verification meets the standard for the kind of story being told.
| Operation | Distinguishing feature | Public output |
|---|---|---|
| AP | Wire-scale C2PA pilots | Wire feed images |
| Reuters | Strict no-compositing standard since 2006 | Wire feed images |
| AFP Fact Check | Volume fact-checking of viral content | Fact-check articles |
| BBC Verify | Public transparency of verification process | Named-verifier on-air analyses |
| Bellingcat | Detailed open-source investigations | Public investigation reports |
| WaPo Visual Forensics | Multi-day reconstruction of single incidents | Long-form forensic articles |
Where C2PA changes the workflow
Credentialed wire imagery is changing the newsroom workflow in specific ways. An image that arrives with a valid C2PA chain through a trusted signer shortcuts several verification steps: the source identity is established cryptographically, the edit history is recorded, and downstream republication can validate against the chain. This reduces the per-image verification labor for credentialed material substantially.
The catch is that credentialed material is still a minority of input. AP, Reuters, and AFP run mixed feeds in which some images are credentialed and most are not. The verification workflow has to handle both cases, which means the full chain of steps is still operational for uncredentialed material and an additional credential-inspection step is added for credentialed material. The aggregate labor is similar to pre-C2PA verification; the labor distribution shifts.
The bigger change is in republication. A credentialed image that another newsroom picks up from a wire can be republished with its credentials intact, and downstream consumers can validate the chain themselves. This pushes some of the verification labor out of the receiving newsroom and into the cryptographic infrastructure. The receiving newsroom's editorial responsibility remains — they still own the publication decision — but the cost of verifying that the wire image is what AP says it is goes down meaningfully.
Project Origin and inter-organizational coordination
Project Origin is the consortium that has done most of the inter-organizational work to make newsroom C2PA practice consistent. The founding members — BBC, CBC, Microsoft, New York Times — have been joined by a growing set of broadcasters, publishers, and technology vendors. The project's published practice notes have informed how individual newsrooms structure their C2PA deployments, and the joint pilots have demonstrated end-to-end credentialed delivery across organizational boundaries.
The IPTC Photo Metadata Standard 2024.1, mentioned earlier on the metadata analysis page, was developed with substantial input from Project Origin participants and represents the editorial-metadata side of the same coordination work. The combination of structured editorial metadata and C2PA cryptographic signing is the working model for how editorial provenance will operate through the rest of the decade.
What it costs
A newsroom verification operation is not free. A dedicated desk at a wire service is staffed by tens to low hundreds of people globally, with infrastructure (tooling, archives, training) on top. The labor is the dominant cost; the technology is incidental. This is part of why the function is concentrated in larger outlets: the fixed cost of building and maintaining a verification operation is hard to justify at small scale, which leaves smaller outlets either underinvested in verification or dependent on the larger wire services to have done the work for them.
The community-funded alternative — Bellingcat's training programs, the EU's WeVerify research project, the Reynolds Journalism Institute's verification fellowships — distributes some of the capability across smaller outlets and independent investigators. This works for specific high-effort cases (Bellingcat's investigations are themselves long-form journalism) but does not provide routine verification capacity at wire-feed scale.
Where the field is moving
The trajectory through 2026 and 2027 is toward C2PA-credentialed input becoming a larger fraction of newsroom inputs, particularly from the major camera and phone manufacturers' integration with editorial workflows. As that share grows, the per-image verification cost drops for credentialed material and stays the same for uncredentialed material. The eventual steady state — credentialed material is a default, uncredentialed material is the marked case — is several years away but is the structural endpoint that most newsroom investment is targeting.
The harder open question is what happens with citizen-submitted and conflict-zone imagery, which has historically been the highest-stakes verification work and which is unlikely to carry credentials anytime soon. The current answer — multi-stage manual verification, geolocation, independent corroboration — continues to be the answer. The C2PA-credentialed future does not solve this category; it just makes the verification of professional material faster, leaving the citizen-submission verification work as the unchanged residual where most of the verification budget will continue to go.