Answers are grounded in this paper and cite their sources.
Bibliometric analysis reveals rapid publication growth and shifts toward generative AI in post-editing research, highlighting persistent gaps in labor and ethics.
This study maps Scopus-indexed research on human post-editing from January 2018 through May 2026, capturing the consolidation of neural machine translation (NMT) and the rise of generative AI. A PRISMA-adapted search identified 1,014 records; screening yielded 292 eligible peer-reviewed articles. Guided by an integrative technology-practice-knowledge framework, the study combines performance analysis, co-authorship and citation-network mapping, keyword co-occurrence analysis, and temporal overlay visualization. Across complete publication years (2018-2025), the annual growth rate is 26.17%, well above benchmarks for established humanities disciplines; the full-span figure (2018-2026) is lower (9.66%) given the partial 2026 count. Publication patterns show an NMT-consolidation phase (2018–2021), a transitional acceleration phase (2022–2023), and an LLM-focused phase (2024–2025). The articles appeared in 146 journals, led by Translation Spaces, the Journal of Specialised Translation, and The Interpreter and Translator Trainer; activity concentrated in Western Europe and East Asia. Keyword analysis reveals seven thematic clusters around two poles: AI tools and industry disruption, and post-editing effort and cognition. Persistent gaps concern ethics, labor, low-resource languages, and longitudinal competence research. The study offers a reproducible baseline and research agenda, bounded by a single database, excluded conference proceedings, partial-year data, and single-reviewer screening.
No takes yet. Share an insight, caveat, or question.
Zakarneh et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: