Editorial illustration: twelve framed portrait silhouettes arranged as entries in a register, one frame outlined in red
Twelve Americans were named on one list in March 2021. One of the twelve publishes this record — the frame in red — which is why the tracker discloses it on every page.

In March 2021 a London nonprofit published a report with a number in it.

The number said twelve people were responsible for up to sixty-five percent of anti-vaccine content on social media. It travelled faster than anything else in the report, because a number is portable in a way an argument is not. Within weeks it was being read from the White House podium. Platforms acted. Twelve Americans became a list, and the list became a citation that other people could use without having to make the case again.

One of the twelve is the founder of GreenMedInfo, a plaintiff in Finn v. Global Engagement Center.

Facebook's own analysts looked at the claim and found the core statistic exaggerated by a factor of at least 1,300.

The report named twelve people. The platforms acted. The number was wrong by three orders of magnitude.

That is the part of this story that is already known, and it is not why we are publishing today. We have spent the past months building something duller and more useful: a public record of every published characterization of those twelve Americans and of GreenMedInfo, 2013 to 2026 — 241 documents, each row linked to its primary source. And the most important findings in it are the ones where we looked for something and did not find it.

Four of those findings correct claims our own side has been making.

Story at a glance

What we built, and why it is boring on purpose

The tracker holds 241 documents. Most of them do not name GreenMedInfo or its founder at all. They cite the list.

That is the mechanism worth seeing. A list, published once, becomes a citation that lets a government agency, a broadcaster, a fact-checker and a platform each act without re-examining the claim underneath. The original number only has to survive long enough to become a reference.

So every row links to its own primary document, and every row carries a severity grade you can disagree with. The point is not to tell you what to conclude. It is to make the chain inspectable, including where it runs out.

The findings where we found nothing

These are ordered from the most defensible to the least.

1. No government payment to CCDH is documented

This is the closest thing here to unassailable, because it rests on a government's own answer. UK written answer 26614, dated 27 January 2025, records that the Department for Science, Innovation and Technology and Ofcom had not provided funding to CCDH in the period addressed.

Beyond that: a search of Contracts Finder returned no UK government contracts or notices naming CCDH. Of the eight grants GrantNav records to CCDH Ltd, every one is from a private foundation and none is from a government funder.

Author's analysis. A negative finding is not proof of absence — it is proof that we looked in the named places and the record was empty. If a government payment exists, it is not in the registers where government payments are recorded, and we would want to see it.

2. No Gates Foundation grant to CCDH

We checked all 41,686 grants in the Gates Foundation's committed-grants database. There is no grant to CCDH.

There is a real Gates adjacency, and it is narrower than what circulates: a Gates-funded Global Citizen series promoted CCDH's chief executive in December 2021. That is a documented promotional relationship with an intermediary. It is not a grant, and it is not evidence that Gates money paid for the Dozen report.

3. Seven trusts did not fund CCDH — the merger misread

Seven named trusts have circulated widely, including in our own earlier work, as disclosed CCDH funders: the STP Ecclestone Foundation, The Genie Trust, the Kohn Foundation, the Edith & Ferdinand Porjes Charitable Trust, The Harbour Charitable Trust, The Vail Foundation and The David Goldman Foundation.

Each appears in the Charity Commission registration history of Prism the Gift Fund under “funds received from.” That phrase does not mean what it looks like. Each entry is a registered merger: the trust transferred its assets to Prism and closed. It is a dissolution record, not a grant record.

The clearest demonstration is chronological. The STP Ecclestone Foundation merged in 2012. CCDH was incorporated in October 2018. A trust that ceased to exist six years earlier did not fund it.

Author's analysis. Call this the merger misread, because it will happen again. A charity register is a legal instrument, not a plain-English document, and “funds received from” in a registration history describes the registrant's own history of absorbing other charities. We got this wrong, several outlets got it wrong, and the error is now the first thing our funder page tells you. We apologise to the seven trusts named.

What is actually in the record

Removing the claims that do not hold leaves a documented record that is narrower and, we think, more serious — because every item is a primary document and none of it requires a payment to exist.

Author's analysis. Proximity is not payment, and we will not pretend otherwise. None of the above is a contract. What it describes is an organisation operating close enough to government that ministers cited its guidance and officials asked it to shape a brief — while, on the record, not being paid by them. That is a harder thing to talk about than a payment, and it is what the documents support.

What this is not

Building the firewall before anyone lights the match:

Our own corrections

Four errors of ours are now logged in the tracker's correction log, which is public and dated:

  1. We listed the seven Prism-merged trusts as disclosed CCDH funders. They are mergers, not grants.
  2. We stated that no payment from the Arabella network to CCDH had been traced. A tax-filed Hopewell Fund grant from 2021 is recorded in the directory.
  3. We counted that same grant twice, once under each of its two labels, inflating our own total.
  4. We described a count of grant entries as a count of funders. Repeat grants from one organisation are separate entries, not separate organisations.

We also corrected a rounded figure to the exact amount in the filing it came from, which moved one of our totals by $245. A $245 error in a nine-figure context is trivial in every way except the one that matters: two of our own pages were showing different numbers, and that is the cheapest possible way to lose an argument.

How this was checked, and what that is worth

Every funding row in the directory was compiled by an AI research assistant working against named primary sources — Form 990s, Charity Commission filings, Companies House records, government registers and funders' own grant databases — and every row carries a link to the document it came from. None of those rows yet carries a second, human reviewer. We are saying so here rather than letting you discover it.

That is why the design of the thing matters more than our assurances about it. You do not have to trust the compiler. Each row names its document, and the documents are public. Where a widely published claim could not be substantiated, the page says so and keeps the row visible under “checked and not substantiated,” rather than quietly deleting it.

Corrections are welcome and will be logged with a date: coalition@restorethe1st.com.

Why publish the findings that weaken our own case

Because a number that nobody checked is what did this to us.

Twelve people were named on the strength of a statistic that the largest platform in the world internally found to be wrong by a factor of at least 1,300, and no institution that acted on it went back to check. The injury was not primarily that someone lied. It was that a citable claim moved faster than anyone's willingness to verify it.

If our answer to that is a mirror image — a funding web with a famous billionaire at the centre, seven trusts that never gave a penny, and a total inflated by a grant we counted twice — then we have not built a record. We have built the same instrument and pointed it the other way.

So the seven trusts are exonerated on our page. The Gates row says no grant found in 41,686 checked. The government row says no payment documented, and cites the government's own answer. What survives that treatment is worth something precisely because it survived it.

The record is public. So is the funder directory. Every row links to its document. Check us.