After 6 months of fact-checking for Facebook, Full Fact published 96 fact-checks, was paid ~$171k, wants Facebook to share more data with fact-checkers
In December 2016, Facebook enlisted a handful of U.S.-based news organizations (ABC News, Snopes, PolitiFact, FactCheck.org, and the AP) …
Context & Ripple Effects
Full Fact's six-month report card lands at a delicate moment for Facebook's fact-checking program: the original US partners have been thinning out, with AP following Snopes in stepping back from fact-checking work for Facebook earlier this year. A UK partner publishing 96 checks — and attaching a price tag of roughly $171,000 — gives the first public look at the economics of the arrangement.
The report's central demand is also the program's oldest sore point: Full Fact wants Facebook to share more data, echoing [[a:922045|the 2017 complaint that Facebook was withholding data from media partners on privacy grounds]]. Facebook has since leaned harder on machine learning and the ClaimReview framework in its false-news updates, which makes human partners' data access the live question.
First-order effects
- Facebook now faces a public, quantified critique from an active partner: Full Fact is still doing the work but says it cannot judge what to check, or how well checks perform, without better data access.
- The disclosed ~$171,000 payment sets a reference point for what Facebook's fact-checking contracts cost, which every current and prospective partner can now benchmark against.
Second-order effects
- With AP and Snopes already gone, Full Fact's data demands become a test case: if Facebook declines, remaining partners have a documented grievance for following them out; if it concedes, the privacy rationale it cited in 2017 weakens for everyone else.
- Facebook's expanding machine-learning and ClaimReview tooling functions as an internal hedge — the more automated the pipeline, the less the program depends on partners who keep threatening to leave.
Third-order effects
- The pattern points toward platform-funded fact-checking consolidating into a smaller, data-constrained set of partners, with the platform — not the checkers — controlling what information flows and therefore what accountability is possible.
- If the data-access standoff repeats across partners and markets, it pressures regulators and standards bodies to treat fact-checkers' data rights as part of platform accountability rather than a private contract term.
The trend: Facebook's fact-checking program is shifting from a broad network of human partners toward machine-led review, with data access — not payment — becoming the deciding factor in whether partners stay.