Comparative Analysis of Human and ChatGPT Translations in English-to-Arabic Contexts

A new study on artificial intelligence has revealed that human translation slightly outperformed ChatGPT translation in English-to-Arabic contexts, highlighting the continued necessity of post-editing with AI translation tools. The research, conducted by a team at October University for Modern Sciences and Arts, compared outputs from human translators with those from ChatGPT 3.5, utilizing a corpus compiled from the "Our Cartoon President" series and its Arabic translations.

Key Takeaways:

  • The study compared the outputs of human translators with those of ChatGPT 3.5 in English-to-Arabic translations, using a corpus of 32 randomly selected examples from the "Our Cartoon President" series.
  • The research found that human translation slightly outperformed ChatGPT translation, as demonstrated by both qualitative and quantitative analyses, with human translators achieving a higher fluency and accuracy score.
  • The study underscored the continued necessity of post-editing, even with the advent of AI translation tools like ChatGPT, to ensure the highest quality and accuracy in translated content.
  • The research emphasized the potential utility of ChatGPT as a subtitling tool to expedite the translation process, leveraging its efficiency and speed while maintaining the accuracy of human translation.
  • The study employed a mixed-methods approach, combining qualitative insights from professors' evaluations with quantitative data from their ratings to assess the effectiveness of human and ChatGPT translations.
  • The research concluded that human translation remains the gold standard for accuracy and fluency, but ChatGPT can serve as a valuable tool for initial translation, allowing for post-editing by human translators to refine the output.

Statistics:

  • 32 examples were selected for comparison between human and ChatGPT translations.
  • 1-5 scale was used to rate fluency and accuracy of translations by English-Arabic translation professors.
  • A T-test was conducted to determine statistically significant differences between human and ChatGPT translations.
  • 100% of professors agreed that human translation achieved higher fluency and accuracy scores compared to ChatGPT translation.
  • 75% of professors considered ChatGPT to be a useful tool for initial translation, with 25% suggesting it as a primary translation tool.

Sources:

  • "Utilizing ChatGPT for subtitling movies: a comparative analysis of accuracy and fluency with human translators." Discover Artificial Intelligence, 2025, 5(1): 1-27. doi: 10.1007/s44163-025-00310-z.
  • October University for Modern Sciences and Arts. Faculty of Languages. 6 October City.
  • NewsRx. "Research on Artificial Intelligence Detailed by a Researcher at October University for Modern Sciences and Arts (Utilizing ChatGPT for subtitling movies: a comparative analysis of accuracy and fluency with human translators)." Robotics & Machine Learning. October 27, 2025; p 705.