Leveraging ESG disclosure analysis using Artificial Intelligence

Resumo

Driven by recent technological innovations, particularly involving Large Language Models (LLMs) and their applications in the business field, this research project proposes the use of Artificial Intelligence to analyze the content of corporate reports from listed companies. The analysis focuses on the Environmental, Social, and Governance (ESG) perspective, a topic that has been prominent on the agenda of regulators, managers, and researchers. Methodologically, the reports' content is analyzed based on a data collection instrument validated by specialized literature, which breaks down the three ESG pillars into 15 topics and 248 keywords. The Google-Gemini Artificial Intelligence model was used in the testing phase, but the developed method, as well as its corresponding prompt, can be adapted for other AI models, such as ChatGPT. It should be noted that the proposed analysis involves open-access corporate reports made publicly available by the respective listed companies.