Greenwashing Detection: What Does a System Miss?

A company’s report can contain plenty of statements about sustainability. But finding an environmental claim and deciding whether it is supported by evidence are two different tasks. That distinction is at the centre of my research on greenwashing.

Working with my mentor and co-author, Dr. Emre Atlıer Olca, I explored a question about automated analysis: when a rule-based system extracts environmental, social and governance (ESG) claims from corporate reports, what does it miss? A rule-based system follows predefined instructions to identify relevant text. It provides a starting point, but its output depends on what those rules can recognise. My paper focuses on examining that baseline and its limitations.

The title is “Measuring What a Rule-Based ESG Claim Extractor Misses: A Baseline and Its Limits for Greenwashing Detection.” The word “misses” matters. Evaluating a system means looking beyond the claims it successfully identifies and examining the relevant material it leaves out.

There is another distinction I want to keep clear: identifying an ESG claim does not establish that a company is greenwashing. Assessing whether a statement is misleading requires further evidence and context. Claim extraction is one part of that larger process.

What interests me about this project is the connection between economics and technology. Corporate reports communicate information that people may use to make decisions. Computational tools can help us examine that information, but we also need to understand the limits of the tools themselves.

My paper has now been accepted for oral presentation at the 10th International Symposium on Multidisciplinary Studies and Innovative Technologies (ISMSIT 2026), taking place on November 13–15 in Antalya, Türkiye. As I prepare for the presentation, I want to explain both the purpose of the system and the boundaries of what it can tell us. For me, that is one of the most interesting parts of research: making a question precise enough to investigate, while being equally precise about what remains unanswered.