Kenyan and Nigerian SMB workers found generative-AI citations routinely unverifiable or nonexistent, forcing after-the-fact evidence reconstruction

generative-ai-at-workcscwhallucinationcitationstraceabilitykenyanigeriasmall-businesssingle-source-cap

In Abdulhamid, Nyairo, Ankrah & O'Neill's interview study of generative-AI use in Kenyan and Nigerian small businesses (arXiv:2610.31344, 2026), participants reported that AI-supplied citations and references were often unverifiable or nonexistent. A sales and marketing lead described clicking cited links and finding they "always goes to 404," concluding: "Your answer is not wrong, but your source... the reference is so wrong, it's nonexistent." A second participant reported warning colleagues that when a tool is asked for references, "some of the references... may be fake" and "you have to find out."

The paper treats this as the central empirical marker of a "traceability breakdown" that release work exists to repair: outputs sounded plausible while arriving with no retrievable evidence trail, so workers had to "go and do research to get the reference points to defend what it has outputted" after the fact, rather than trusting the tool's own citations at the point of generation. The paper's own design recommendation drawn from this finding warns against building citation features that merely "manufacture confidence" without fixing the underlying grounding gap — a distinction between a feature that looks like verification and one that actually provides it.