Combining Esg Ratings and Media Content For A Better Understanding of Corporate Ethics: An Ai Perspective

This paper combines two artificial intelligence tools - topic modelling (LDA) and sentiment analysis - to analyse news articles about European companies, focusing on ethical/unethical behaviours. The aim is to complement traditional ESG ratings by identifying discrepancies between ESGC scores and media content. Findings reveal that companies with high volumes of news tend to score lower in ESGC rankings. Although, in general, findings indicate a strong correspondence between ESGC scores and textual evidence, they also reveal gaps where the sentiments and topics identified does not fully align with the scores. The research highlights the importance of combining ESG scores with new media's thematic and emotional insights to better identify ethical business practices. This novel framework provides a deeper understanding of ESG performance and allows managers to align strategies with market perceptions. Overall, the main contribution is to enhance the transparency and reliability of ESG assessments for stakeholders and investors.