Scholarlink

<p>Media play an essential role in shaping public attitudes, behaviors, and long-term societal developments. Although AI is often addressed in the context of technical advancement, its role in media discourse on Sustainable Development Goal 1 (SDG1 - No Poverty) is underexplored. Media sentiment towards AI and poverty is researched here through sentiment analysis, topic modeling, and predictive modeling of global news headlines on the GDELT database. This study addresses this gap by examining worldwide media sentiment and narrative trends on AI and poverty from 2017 to 2025. Sentiment classification used VADER, TextBlob, and BERT models, with a net positive sentiment reported in AI-poverty stories. Latent Dirichlet Allocation (LDA) topic modeling revealed labor workforce automation and sustainable economic transformation to be the most discussed themes. In addition, Long Short-Term Memory (LSTM) neural networks were employed to predict sentiment trends. The model demonstrated low predictive validity (R² = 0.0124; Pearson correlation = 0.1223), suggesting that additional contextual variables are needed. Notably, BERT performed a more accurate detection of subtle changes in sentiment than VADER and TextBlob, which tended to make more coarse categorizations. The LDA model further revealed consistent discussions surrounding government policy, labor challenges, and ethical concerns about automation. These findings highlight the polarized nature of AI-related media coverage in poverty contexts and the value of sentiment analysis in assessing AI’s perceived socio-economic role. Our study contributes to the literature by combining computational approaches to evaluate sustainability-oriented media narratives.</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

Sentiment analysis of news: unveiling AI’s role in sustainability and no poverty (SDG1)

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年7月1日

查看全文
  • Robin Kunju Mol Raj,
  • Marek Vochozka,
  • S Sujatha

摘要

Media play an essential role in shaping public attitudes, behaviors, and long-term societal developments. Although AI is often addressed in the context of technical advancement, its role in media discourse on Sustainable Development Goal 1 (SDG1 - No Poverty) is underexplored. Media sentiment towards AI and poverty is researched here through sentiment analysis, topic modeling, and predictive modeling of global news headlines on the GDELT database. This study addresses this gap by examining worldwide media sentiment and narrative trends on AI and poverty from 2017 to 2025. Sentiment classification used VADER, TextBlob, and BERT models, with a net positive sentiment reported in AI-poverty stories. Latent Dirichlet Allocation (LDA) topic modeling revealed labor workforce automation and sustainable economic transformation to be the most discussed themes. In addition, Long Short-Term Memory (LSTM) neural networks were employed to predict sentiment trends. The model demonstrated low predictive validity (R² = 0.0124; Pearson correlation = 0.1223), suggesting that additional contextual variables are needed. Notably, BERT performed a more accurate detection of subtle changes in sentiment than VADER and TextBlob, which tended to make more coarse categorizations. The LDA model further revealed consistent discussions surrounding government policy, labor challenges, and ethical concerns about automation. These findings highlight the polarized nature of AI-related media coverage in poverty contexts and the value of sentiment analysis in assessing AI’s perceived socio-economic role. Our study contributes to the literature by combining computational approaches to evaluate sustainability-oriented media narratives.