Verification under deadline: the core challenge
Breaking news creates intense pressure to publish quickly, but speed without verification causes serious harm — to individuals falsely named, to the public misled by false information, and to the credibility of the outlet that published it. The social media age has amplified both the volume of unverified claims and the speed at which false information spreads.
UK journalism standards require accuracy as the primary obligation. IPSO's Editors' Code Clause 1 states that “the press must take care not to publish inaccurate, misleading or distorted information or images.” There is no deadline exception. A systematic verification workflow is not a luxury — it is the professional standard.
This guide covers the key verification tools and techniques, how to assess source credibility quickly, how to identify synthetic or recycled media, and when it is appropriate to publish with caveats rather than holding the story entirely.
When verification workflows are most critical
- 1A major breaking-news event with rapidly circulating social media claims and no official confirmation.
- 2A viral image or video that appears to document a newsworthy incident but cannot be traced to a known source.
- 3A claim from an anonymous or pseudonymous social media account about a named individual or organisation.
- 4A tip from a source who has been accurate in the past but whose current claim is unusually dramatic or convenient.
- 5A document, screenshot, or recording that a source says they obtained from inside an organisation.
- 6Any claim involving AI-generated content, manipulated imagery, or synthetic audio.
- 7Breaking news in a conflict zone or during a civil emergency where misinformation is deliberately weaponised.
Red flags to watch for
- Single-source social media claims: a claim circulating widely on social media does not become verified by volume — it becomes viral misinformation.
- Recycled imagery: a reverse image search returns older results showing the image predates the claimed event.
- Synthetic media signals: AI-generated images often show characteristic artefacts around hands, text within images, and background details.
- Suspiciously perfect timing: a leaked document or image that arrives exactly when it would most harm a particular target should raise your verification threshold.
- Source incentive: a source who has a strong personal, financial, or political reason for the claim to be true needs independent corroboration.
- Confirmation bias: a claim that perfectly matches the narrative you already believe about a subject deserves extra scepticism, not less.
- Anonymous mass-emailing: a document or claim distributed simultaneously to many journalists, without any apparent source, should be verified as primary material, not published on the basis of the email alone.
Breaking-news verification checklist
- I have identified at least two independent sources with direct knowledge of the claim (not two sources citing the same origin).
- I have run any key images or video through a reverse image search (Google Images, Yandex, TinEye, or InVID/WeVerify).
- I have checked the earliest known upload of any video or image and confirmed it matches the claimed date and location.
- I have attempted geolocation of visual material using landmarks, street furniture, mapping tools, and satellite imagery.
- I have checked whether any named individuals can be contacted directly for comment before publication.
- I have assessed whether any documents could be fabricated, and cross-referenced their details against publicly available records.
- I have considered whether AI-generated or deepfake media could be involved, and applied appropriate detection techniques.
- I have consulted my editor on any claim I cannot verify but believe carries significant public interest.
- I have ensured any caveats about unverified claims are prominently worded in the published piece, not buried.
- I have kept a verification log documenting what I checked, when, and what the result was.
Tools for verification and social media checking
Use our Social Media Verifier checklist to work through a systematic verification process, and our Investigative Reporter Pack for document verification templates.
Common mistakes
- Treating high social media engagement as a proxy for accuracy — false information often spreads faster than true information.
- Verifying the source of a claim rather than the claim itself — even a reliable source can be wrong or deceived.
- Not running reverse image searches because the image “looks real” — synthetic images increasingly do.
- Failing to update or correct a breaking-news story when subsequent verification reveals an error.
- Using “reportedly” or “allegedly” as a substitute for verification rather than as a transparency marker alongside genuine verification.
- Not keeping a verification log: if you are challenged later, your methodology record is your primary defence.
- Ignoring geolocation checks on images: visual verification is among the most reliable forms of verification for imagery.
Related guides
Primary sources
- Bellingcat — Verification Methodology Guides— Bellingcat
- Information Futures Lab (formerly First Draft)— Brown University
- BBC Verify — How We Check Claims— BBC
- IPSO Editors' Code of Practice — Clause 1 Accuracy— IPSO
- Amnesty International Citizen Evidence Lab— Amnesty International
- InVID / WeVerify Browser Extension— WeVerify Project
Frequently asked questions
What is the minimum number of independent sources needed before publishing a breaking-news claim?
There is no universal rule, but the standard editorial practice at major UK outlets is at least two independent sources who have direct knowledge of the claim — not two sources who both received the same information from the same origin. “Independent” means they have separate, first-hand access to the facts. For serious allegations, three sources is a safer threshold. If only one source exists, you need compelling corroborating documentation before publication.
How do I verify a social media video claim during breaking news?
Work through these steps: (1) Run the video through a reverse image search tool (InVID/WeVerify browser extension is the standard tool for this). (2) Check the metadata if accessible — original upload date, device model, GPS coordinates embedded in EXIF data. (3) Geolocate the scene using visible landmarks, street furniture, signage, or Google Street View. (4) Check whether the audio matches the claimed location (language, accents, emergency service tones). (5) Search for the earliest known upload of the same clip to identify whether it was captured at the time claimed.
What are the main signs of AI-generated or deepfake imagery?
Common indicators include: unnatural skin texture particularly around the hairline; inconsistent lighting or shadows that do not match the claimed environment; blurred or distorted background details (especially text and hands in earlier-generation models); ear shapes that are asymmetric or irregular; and metadata that shows the image was created by generative software. Tools to assist detection include Google’s About this Image feature, Hive Moderation’s deepfake detector, and the C2PA Content Credentials standard (where embedded). No tool is definitive — multiple signals together build the case.
What is the First Draft methodology for breaking-news verification?
First Draft (now part of Information Futures Lab) developed a widely adopted breaking-news verification framework. Core steps are: (1) Stop — do not amplify until verified. (2) Investigate the source of the claim. (3) Find the original. (4) Read the full content, not just the headline. (5) Check the date of original publication. (6) Ask an expert. (7) Ask yourself “is this too good to be true?”. The Information Futures Lab publishes updated methodology guides at informationfutures.org.
How should I handle a tip from a social media account I cannot verify?
Treat unverified social media accounts as leads, not sources. Check the account’s creation date, follower count, posting history, and whether it is associated with any real-world entity. Do not attribute claims to unverified accounts as if they are reliable sources. If the account has made accurate claims in the past, that is a factor but not sufficient on its own. Use the tip to direct further reporting — can you verify the underlying claim through documents, officials, or eyewitnesses?
When should I publish an unverified claim and label it as such?
In rare circumstances, when a claim is so significant that the public interest requires informing the audience even before verification is complete, you can publish with clear labelling: “UK JournoHub has not independently verified…” or “This claim has not been confirmed by…”. This approach carries real risk — it can amplify misinformation, cause harm to innocent individuals, and damage your credibility. It should be used only with editorial sign-off at a senior level and only when the story is significant enough that the public genuinely needs to know even the unverified claim.