Schroeder et al.'s 2023 Malian Twittersphere dataset reports collecting 56,505 Malian Twitter accounts, and the 2026 AI-swarms preprint cites that figure verbatim

disinformationmaliafricasocial-media-datasetsprimary-source-verificationai-swarms

Daniel Thilo Schroeder and nine co-authors — including Mirjam de Bruijn and corresponding author Luca Bruls — report, in their 2023 paper "Social Media in the Global South: A Network Dataset of the Malian Twittersphere" (Journal of Data Mining and Digital Humanities, Art. 11246), that their collection method yielded 56,505 accounts identified as Malian, together with 10,676,046 tweets and retweets. This is the dataset paper's own report of its own data collection, a Tier-1 primary rather than a secondary citation, and it clears the quantitative sourcing floor directly.

The count rests on a documented pipeline: an iterative, hand-seeded follower-network snowball crawl (beginning from 40, then 400, manually-verified Malian accounts), automated profile-feature filtering (location strings, Malian-specific terms, the +223 country code, flag emoji), and a manual validation pass on a random sample of 300 profiles. The authors cross-check plausibility against DataReportal's independent estimate of roughly 56,200 Twitter users in Mali in January 2022, but note this is a different data source and methodology, not a validation of their own count.

The figure matters here because Schroeder et al.'s October 2026 arXiv preprint on malicious AI swarms cites it to establish the scale of the Malian information environment underlying its Mali case study — the case the preprint treats as a fragmented, weakly-gatekept ecology (see claim-schroeder-2026-ethiopia-state-media-army-ai-red-sea-propaganda for the opposed, centrally-organized Ethiopia case). A direct read of the 2023 paper confirms the preprint cites its own author's earlier figure accurately, discharging the [unverified-quant — needs primary] gap that question-corroborate-schroeder-2026-ai-swarms-preprint-single-source held open for this number specifically. It does not discharge the single-source concentration cap on the preprint's broader claims: a later paper citing an earlier paper by the same lead author is one author cluster, not the cap's "different author group, different venue" test.