Detecting greenwashing signals through a comparison of ESG reports and public media
In 2022, greenwashing is at the top of the social agenda by companies, and there is no sign of relinquishing its position. Greenwashing happens when companies over-report positive data about their sustainability efforts while downplaying the negative impacts of their operations. The reason for this is the subjective nature of most ESG (Environmental, Social, and Governance) information. Public companies are required to report on their sustainability efforts. However, when investors and ESG rating providers focus on self-reported data (ESG reports and marketing communications), they often receive a distorted and overly positive picture of the company’s ESG performance. Because of this information asymmetry, even investors who want to invest sustainably can be misled in their investment decisions.
Data from external media can help address this problem. To a large extent, third-party content
providers have no interest in promoting a particular company’s sustainability efforts – so by
considering data from a variety of media outlets, we can form a more objective and even critical
picture of a company.
To get an overview of greenwashing, you can consult this website:
https://www.clientearth.org/projects/the-greenwashing-files/.
Aim
In this task, we want to:
-
- Better understand the nature of greenwashing through large-scale text analysis
- Investigate whether sentiment alalysis can shed light into greenwashing risks
- Analyse whether specific ESG topics (for example, currently “trending” topics) are more prone
to greenwashing
Data and Registration
Please find all information related to data access and registration under the organiser’s website: https://sites.google.com/view/greenwashingswisstext/home
Task
To conduct this task, we provide participants with a comprehensive ESG dataset covering company ESG reports as well as public media targeting a wide range of different stakeholders (including investors, NGOs, regulators, society, etc.). The task is to develop approaches to identify gaps and inconsistencies between company-reported data and “external” data that may indicate greenwashing.
This is an application-oriented task. While there are no specific requirements for the NLP algorithms to be used, we suggest to focus on the following three aspects:
-
- Greenwashing takes place at the information level [1]. Using a large dataset, we want to better understand the nature of greenwashing and try to quantify the degree.
- We want to prototype NLP approaches to detect greenwashing using public documents reflecting different stakeholders.
- We want to visualize potential indicators of greenwashing as well as the reliability of these indicators in a credible way.
Important dates
- April 10th: Organizers release data sample. (8 weeks before the actual workshop)
- June 1st: Teams submit test set results.
- June 12th: Workshop day
Participant
General audience and Bachelor/Master students at Universities (of Applied Sciences).
References
[1] Naderer, Brigitte, Desirée Schmuck, and Jörg Matthes. “‘2.3 Greenwashing: Disinformation
through Green Advertising.” Commercial communication in the digital age: Information or
disinformation 105 (2017): 120.
