Scholarlink

<p>This study presents a large-scale computational mapping of global public discourse on generative AI, focusing on ChatGPT. Drawing on GDELT metadata, the study examines 141,509 news articles across more than 100 languages published during the two years after ChatGPT’s launch to map changes in thematic framing, sentiment, and geographic diffusion. The analysis identifies a clear lifecycle in discourse volume, transitioning from initial novelty-driven coverage to stabilized, institutional attention. Thematic patterns reveal a strong emphasis on education, professional roles, governance, and sociocultural inclusion. Sentiment analysis shows a general trend toward cautious optimism, but with significant regional asymmetry: news coverage in South and East Asia exhibits higher affective positivity, while Western media shows greater regulatory skepticism. Cross-linguistic sentiment mapping further reveals distinct emotional resonances, with Hindi, Urdu, and Spanish-language sources expressing higher optimism than Arabic and Russian counterparts. The study demonstrates how large-scale computational discourse analysis can provide a foundational empirical map of a technology’s entry into global public consciousness, offering critical insights for policymakers and scholars navigating the societal integration of AI.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A computational mapping of global digital discourse on ChatGPT and the emergence of generative artificial intelligence

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月6日

查看全文
  • Muhammad Bilal Zafar

摘要

This study presents a large-scale computational mapping of global public discourse on generative AI, focusing on ChatGPT. Drawing on GDELT metadata, the study examines 141,509 news articles across more than 100 languages published during the two years after ChatGPT’s launch to map changes in thematic framing, sentiment, and geographic diffusion. The analysis identifies a clear lifecycle in discourse volume, transitioning from initial novelty-driven coverage to stabilized, institutional attention. Thematic patterns reveal a strong emphasis on education, professional roles, governance, and sociocultural inclusion. Sentiment analysis shows a general trend toward cautious optimism, but with significant regional asymmetry: news coverage in South and East Asia exhibits higher affective positivity, while Western media shows greater regulatory skepticism. Cross-linguistic sentiment mapping further reveals distinct emotional resonances, with Hindi, Urdu, and Spanish-language sources expressing higher optimism than Arabic and Russian counterparts. The study demonstrates how large-scale computational discourse analysis can provide a foundational empirical map of a technology’s entry into global public consciousness, offering critical insights for policymakers and scholars navigating the societal integration of AI.