[Paper Review] Artificial intelligence contribution to translation industry: looking back and forward

Mohammed Q. Shormani, Al-Sohbani, Yehia A.|arXiv (Cornell University)|Nov 29, 2024

Translation Studies and PracticesArts and Humanities3 citations

TL;DR

This study conducts a comprehensive scientometric and thematic analysis of 9,836 unique AI-related translation research articles (1980–2024) from WoS, Scopus, and Lens, revealing that neural machine translation and large language models like ChatGPT have significantly advanced the field. Despite progress, critical challenges remain in low-resource, multi-dialectal, and culturally nuanced language translation, highlighting the need for further rigorous research in these areas.

ABSTRACT

This study provides a comprehensive analysis of artificial intelligence (AI) contribution to research in the translation industry (ACTI), synthesizing it over forty-five years from 1980-2024. 13220 articles were retrieved from three sources, namely WoS, Scopus, and Lens; 9836 were unique records, which were used for the analysis. We provided two types of analysis, viz., scientometric and thematic, focusing on Cluster, Subject categories, Keywords, Bursts, Centrality and Research Centers as for the former. For the latter, we provided a thematic review for 18 articles, selected purposefully from the articles involved, centering on purpose, approach, findings, and contribution to ACTI future directions. This study is significant for its valuable contribution to ACTI knowledge production over 45 years, emphasizing several trending issues and hotspots including Machine translation, Statistical machine translation, Low-resource language, Large language model, Arabic dialects, Translation quality, and Neural machine translation. The findings reveal that the more AI develops, the more it contributes to translation industry, as Neural Networking Algorithms have been incorporated and Deep Language Learning Models like ChatGPT have been launched. However, much rigorous research is still needed to overcome several problems encountering translation industry, specifically concerning low-resource, multi-dialectical and free word order languages, and cultural and religious registers.

Motivation & Objective

• To analyze the evolution and impact of artificial intelligence on the translation industry over 45 years (1980–2024).
• To identify key research trends, hotspots, and emerging themes in AI-driven translation through scientometric and thematic analysis.
• To assess the contributions of neural networks and large language models such as ChatGPT to translation quality and system capabilities.
• To highlight persistent challenges in translating low-resource languages, Arabic dialects, and culturally or religiously sensitive texts.
• To guide future research by identifying gaps and priorities in AI for translation, especially in complex linguistic and cultural contexts.

Proposed method

• Conducted a systematic literature search across three databases—WoS, Scopus, and Lens—yielding 13,220 articles, of which 9,836 were unique records.
• Performed scientometric analysis using metrics including cluster distribution, subject categories, keyword bursts, centrality scores, and research center contributions.
• Applied thematic analysis to 18 purposefully selected articles to examine their purpose, methodology, findings, and contributions to future AI translation research.
• Tracked keyword bursts to identify emerging and declining research topics over time, such as 'neural machine translation' and 'large language model'.
• Mapped research output by country and institution to assess global contributions and collaboration trends in AI for translation.
• Used citation and co-occurrence analysis to evaluate the influence and connectivity of key research themes and leading institutions.

Experimental results

Research questions

• RQ1 How has the contribution of artificial intelligence to the translation industry evolved from 1980 to 2024 in terms of research output and thematic focus?
• RQ2 What are the dominant research themes, keywords, and emerging trends in AI for translation, particularly concerning neural networks and large language models?
• RQ3 Which linguistic challenges—such as low-resource languages, Arabic dialects, and free word order—remain under-researched despite AI advancements?
• RQ4 How do centrality and keyword burst analysis reflect shifts in research priorities over time in AI-driven translation?
• RQ5 What are the key gaps in current AI translation research, especially regarding cultural and religious register translation, and what future directions are suggested?

Key findings

• Neural machine translation and large language models such as ChatGPT have become dominant forces in advancing AI's role in translation, significantly improving system performance and usability.
• Keyword burst analysis identified 'neural machine translation' and 'large language model' as major rising trends, indicating growing research focus since the 2010s.
• Despite progress, research on low-resource languages, multi-dialectal Arabic, and culturally or religiously sensitive texts remains insufficient, indicating critical knowledge gaps.
• The most influential research centers and authors are concentrated in North America, Europe, and East Asia, with limited representation from the Middle East and Africa.
• Subject category analysis revealed that 'Artificial Intelligence' and 'Computer Science' are the primary domains driving AI translation research, with increasing interdisciplinary overlap.
• Thematic analysis of 18 selected articles confirmed that while AI enhances translation speed and scalability, challenges in semantic accuracy, cultural nuance, and register preservation persist.