Mapping the Structure of Artificial Intelligence in Accounting

Research on artificial intelligence (AI) in accounting has expanded rapidly, yet bibliometric reviews report different corpora and structures. This study examines the field‘s development, conceptual and intellectual structures, and underrepresented research concerns. A rerunnable Web of Science Core Collection search and screening yielded 610 documents published from 1982 to 2025. Methods included performance, co-word and co-citation analyses, pairwise removal of four frequent terms—artificial intelligence, accounting, auditing, and machine learning—and reference publication year spectroscopy. Two applications of the screening protocol agreed on 93.5% of 3,220 records; cases in which both departed from the recorded decision were adjudicated by the lead author. This study finds that five keyword clusters were organised mainly by AI technique rather than accounting task, while pairwise removals involving artificial intelligence, accounting, and auditing produced the largest structural changes. The cited base combined accounting research, analytical methods, and work on automation and judgment. Generative-AI terms appeared in established accounting applications but had few direct links to earlier technology terms, while keywords for external validation, robustness, and governance remained uncommon. The study contributes a rerunnable search, record-level screening decisions, and a sensitivity-tested interpretation of the field‘s bibliometric structure.