Same Structure, Different Visibility: Productivity, Persistence and Citation Visibility of AI Authors in Iran and Four Peer Systems, 2010–2017

Purpose. Rapid growth of a national research field raises questions about its human base: is output concentrated in a few prolific authors, do newcomers persist, and is productivity matched by visibility? Evidence for Iranian artificial intelligence (AI) is missing, and so is a comparison with systems of similar size and development. Design and methods. Using OpenAlex records of all AI articles published in 2010–2017 with an author affiliated in Iran, Türkiye, Egypt, Pakistan or Malaysia (16,467 authors), we compare the distribution of papers per author (Gini coefficient, top-decile share, Lotka-type exponent), persistence (publishing in several years, retention of 2010–2012 entrants to 2015–2017), renewal of the author base, and the relation between productivity and citation visibility, with bootstrap intervals, permutation tests and regression with productivity-class and entry-year fixed effects. Findings. The productivity structure is almost identical across systems: about three quarters of authors have a single paper in the window (Iran 74.3%, 95% CI 73.3–75.3; Türkiye 73.7%; Egypt 76.3%; Pakistan 76.1%; Malaysia 72.3%), the Gini coefficient is 0.30–0.37 (Iran 0.363), the top 10% of authors write 32–37% of paper slots, and about a quarter of 2010–2012 entrants still publish in 2015–2017 (Iran 25.5%; Türkiye 24.7%). Iran's Gini did not differ from Türkiye's or Egypt's (permutation p = 0.39 and 0.55) and exceeded Pakistan's (0.363 vs 0.304; p = 0.009). Iran's active author base grew 3.2-fold (539 to 1,710) and the new-author share fell from 86% to 65%. The difference is in visibility: at every productivity level Iranian authors' papers were less likely to be in the citation top-10% than Turkish (adjusted gap 8.7 percentage points, 95% CI 6.9–10.5) and Pakistani authors' (9.1; 6.6–11.7), by about the same as Egyptian authors' (0.5; −1.5 to 2.6), and the gap persisted with entry-year controls (pooled 6.7 pp; 5.3–8.1). Originality. The study separates two things that national assessments often conflate, the structure of the author base and its visibility, and shows that Iran's gap lies in visibility, not in the concentration or persistence of authorship.