By Atiyah Curmally
There is growing interest in tapping Artificial Intelligence (AI) to generate insights on the environmental, social, and governance (ESG) performance of companies, projects, even countries. Over 80 percent of the world’s largest firms by revenue are reporting on sustainability aspects of their businesses and 90 percent of S&P 500 companies disclose ESG data. Credit rating agencies who once ranked companies purely on financial performance are adding ESG metrics to their offerings while new rating firms are now popping up that specialize in ESG rankings.
The United Nations has created the Sustainable Stock Exchanges Initiative that promotes networking and research among capital market stakeholders. Regulators are developing standards with the goal of making ESG ranking methodologies more transparent, which should make it easier for non-specialists to understand and compare ESG data sets. ESG was a central consideration in every individual investment from the more than $30 billion that IFC committed to emerging markets over the past year.
IFC started building its own tool MALENA* back in 2019 as an innovative way to use AI to help map out the ESG landscape for emerging markets. Step one involves IFC’s ESG experts training MALENA to analyze the language contained in documents in a way that will generate the most useful ESG insights. Human language, as we know, is complex and nuanced. Understanding the context of what is written is key to grasping intended meaning — and preventing errors in interpretation.
For example, one could program an AI tool to identify all references to ‘waste’ from a batch of documents. Unless trained otherwise, the tool might consider these references to be negative indicators of environmental performance. But what if the word ‘waste’ was immediately followed by ‘management’? In that context, it may be describing a company’s efforts to manage waste in an environmentally sustainable manner. A lot of time and effort is being paid to such nuances in order for MALENA to become a genuinely useful and accurate tool.
The Making of MALENA
Robust training and testing conducted by IFC enables AI to predict ESG outcomes accurately.
So far MALENA has been trained to identify more than 1,200 ESG-related terms in unstructured text. Currently, it can read only English but other languages will be added to its repertoire as it continues to be refined for better performance and usability. MALENA is based on a technology called natural language processing. First developed in the 1950s, the technology has become much more sophisticated in recent years, thanks to advances made by large tech firms like Google and Meta that also open-source their models.
MALENA has analyzed over 200,000 internal IFC and public documents — mainly impact assessments, news articles, and sustainability reports — which stretch back decades and cover over 20,000 projects in 186 countries. One of MALENA’s big strengths is speed: it reads 19,000 sentences in a minute. After scouring the documents, MALENA generates dashboards on ESG performance. These dashboards can be a helpful reference source for IFC’s ESG experts as they make their risk assessments.
AI can perform fast filtering and analysis of massive data, generating insights that enable more accurate risk assessments by experts.
Although conceived as an inhouse tool, consideration is being given to wider use after its official rollout in 2023. For instance, IFC recently teamed up with Amundi Asset Management, Europe’s largest asset manager, in a project where Amundi tested MALENA on a sample of their documents associated with 804 financial institution issuers of hard-currency debt in 60 emerging markets. The results were promising; MALENA was able to validate Amundi’s in-house ESG scores but perhaps even more significantly, it added information for 29 percent of test samples (236 of 804 issuers) for which Amundi had no scores previously.
With demand for ESG data growing exponentially, including in emerging markets where such data is often lacking, AI tools like MALENA have an important role to play in improving data quality. They can help investors meet their ESG goals in innovative and nontraditional ways.
Published in September 2022