The pitch is polished. The bio sounds plausible. The contributor has clips, a headshot, and a tidy opinion about the exact issue your audience follows.
Fine. Spend 20 minutes trying to prove the person exists.
On October 8, OpenAI reported two influence operations built around false-front organizations and fake contributor identities. One operation used seven fabricated journalist personas to pitch long-form articles to small and midsize publications. OpenAI identified nearly 100 published or syndicated articles across roughly a dozen outlets.
The point is not that every unfamiliar freelancer is suspicious. Most are trying to make a living. The point is that a clean pitch and fluent draft no longer carry much identity weight. Editors, associations, nonprofits, agencies, and company newsrooms need a small verification step before a borrowed byline becomes their problem.
A good article can survive a basic identity check. A fake contributor is counting on you skipping it.
Minutes 0 to 5: verify the person outside the pitch
Start fresh. Do not use only the links the sender supplied.
- Search the name in quotation marks with the claimed employer, beat, city, and recent article titles.
- Open the staff page or contributor archive on the publisher's own domain. A social profile that says someone works there is not the same thing.
- Check whether older clips existed before the current news cycle. Ten articles published last month under one new identity deserve a closer look.
- Compare the bio, location, job history, headshot, and contact details across sources. Record mismatches. Do not explain them away for the sender.
- Search the profile photo with a reverse-image tool when the identity is unclear. Stock imagery, a different name, or a face reused across unrelated profiles ends the quick approval path.
OpenAI found social accounts that presented Western-seeming journalist identities while platform transparency settings pointed elsewhere. One fake research organization claimed hundreds of staff on LinkedIn but listed only two employees. None of those details proves bad intent alone. Together, they turn a routine submission into a verification assignment.
Minutes 5 to 10: make the email address do some work
A Gmail address is common among freelancers. It also tells you almost nothing. Ask for one piece of independent confirmation tied to the claimed identity.
- Reply through the professional address listed on an established publication, employer, university, or organization page.
- Call a publicly listed office number and ask to be connected. Do not use the number in the pitch signature.
- If the contributor claims an editor or colleague, contact that person through a separately verified channel.
- For a first-time contributor, schedule a brief video or phone conversation and ask specific questions about the draft, sources, and reporting process.
A real person may be slow, private, or lightly documented. That calls for more reporting, not an accusation. Keep the language plain: "We verify new contributor identities before publication. Please confirm through one of these channels." Good process should not require amateur spy theater.
Minutes 10 to 15: inspect the work, not the prose
Fluent writing is cheap evidence. Check the reporting underneath it.
- Open every cited source. Does it support the sentence, date, number, and scope?
- Ask for original notes. For interviews, request contact information, dates, recordings or transcripts when appropriate, and permission terms.
- Check the quote trail. Confirm a sample of quotes with the named source through an independent channel.
- Look for recycled framing. Search distinctive phrases. A polished draft may be copied, lightly altered, or assembled from real reporting done elsewhere.
- Name the interest. Ask who commissioned, edited, paid for, or helped place the article and whether anyone mentioned in it has a financial or political relationship with the contributor.
The Iranian operation in OpenAI's report tailored drafts to publication guidelines and generated pitch emails after the articles were polished. That is the uncomfortable part. A manuscript can follow your style sheet and still arrive through a false identity.
Use the evidence habits in SigServe's four-pass creative review: put the proof beside the claim. For AI-assisted submissions, the content provenance policy gives you a simple record for sources, edits, approvals, and published versions.
Minutes 15 to 20: choose a lane and document it
Do not make one nervous editor carry the decision alone. Use three lanes.
Lane one: verified
The identity checks out, the sources support the draft, conflicts are disclosed, and the contributor can answer detailed questions. Continue through normal editing. Save the verification note with the assignment.
Lane two: unresolved
The identity may be real, but the work history, contact channel, source trail, or commissioning relationship remains unclear. Pause publication. Ask a second editor or communications lead to review the evidence. Set a deadline. Silence does not turn an unresolved pitch into an approved one.
Lane three: deceptive
The person or organization is fabricated, source materials are false, or the sender lied about who funded or placed the piece. Preserve the pitch, headers, draft, links, identities, and verification steps. Alert the publisher, counsel, security team, or platform that needs the record. If the article already ran, correct or retract it clearly. Quietly deleting the page leaves readers and syndication partners with no useful explanation.
Build the check into the intake form
Editors should not reconstruct this workflow after a suspicious submission lands. Add a short contributor record to the normal intake process:
- legal name and public byline;
- current professional affiliation and independently verified contact;
- three relevant clips with publication dates;
- commissioning, payment, client, political, and commercial relationships;
- interview and source documentation;
- AI assistance used in research, drafting, translation, images, or editing;
- identity checker, source checker, editor, and approval date.
This does not need to become a 14-page gate that scares off useful contributors. One page. One owner. A second set of eyes when the facts wobble.
PR teams should run the check in reverse
A fake journalist can approach your organization for a quote just as easily as a fake contributor can approach an editor. Before an executive answers an unfamiliar request, verify the outlet, byline, email, assignment, deadline, and intended use. Call the newsroom through its public number when the request is sensitive.
Then brief the spokesperson with the verified context. Our media pitch guide covers the other side of that exchange: a clear, specific, credible approach that gives an editor something worth opening. The same signals help legitimate reporters establish who they are.
If someone is already publishing or buying ads under your organization's identity, move to the 60-minute impersonation response drill. Contributor screening catches a pitch before publication. Incident response starts after the false identity reaches the public.
Keep the friction where it belongs
The fake pitch should absorb the extra work, not every freelancer who writes to you. Publish the requirements. Use the same questions. Explain why a submission is paused. Give legitimate contributors a clear path to verify themselves.
Twenty minutes will not expose every organized influence campaign. It will catch thin identities, borrowed clips, unverifiable sources, hidden clients, and people who disappear when asked to confirm the basics. That is a useful start.
Does your editorial intake process verify the person behind the pitch?
SigServe can tighten contributor screening, source checks, approval roles, and correction plans without turning every submission into a background investigation.