Appropriate use of AI can contribute to the common effort to do well and to do good at the same time.
by Terence Tse and Mark Esposito
Climate change is one of the hottest topics (pun intended) in the media today. But the well-being of our future – and that of future generations – is not just about environmental issues. In addition to the planet, we should also care about people. In 2013, the collapse of the Rana Plaza garment factory building in Dhaka, Bangladesh reminded us of the dire working conditions certain companies subject their workforce to.
Sadly, these malpractices continue into this decade and are not limited to developing countries. Consider the example of the British company Boohoo. In 2020, this ultra-fast retail pioneer was found to be sourcing from a Leicester factory where workers were being paid just £3.50 an hour, well below the then national wage of £8.72. Furthermore, the workers did not have adequate protective equipment against Covid-19.
Earning with capital
Fortunately, interest in achieving environmental, social and corporate governance (ESG) goals is growing exponentially. At the same time, investors and capital holders are rapidly seeing ESG investing as an important new asset class.
According to an analysis by Bloomberg, ESG-related investments are projected to exceed $53 trillion by 2025, accounting for a third of all assets under management globally. In 2020, a survey conducted by investment bank BNY Mellon and the Official Monetary and Financial Institutions Forum found that over three-quarters (77%) of public investors globally have already implemented ESG in their investment processes . These are important developments because what is best for the planet and for people can now lead to greater profits. We are more likely to achieve sustainability goals when they are aligned with investor interests.
Despite the growing enthusiasm, however, a major obstacle remains for financial firms providing ESG products and services: the lack of timely and accurate data. For example, before its crash in June 2020, German payment processor Wirecard had received average ratings from several ESG rating agencies, even though stories about its questionable business practices had begun to emerge in 2015. Before its unethical business practices were disclosed, Boohoo had received a double-A ESG rating from an agency – the agency's second highest – and had an average score far above the industry on the chain's labor standards. supply.
How could ESG ratings have been so flawed? One reason is that assessments of the ESG quality of companies often rely heavily on information provided by the companies reviewed. This is equivalent to giving the latter the green light to choose which data to use.
Another reason is that the latest news and information is not collected or incorporated in time. To solve this problem, a lot of resources and manpower must be spent on searching, collecting, processing, entering and analyzing data. Until now, Internet news gathering and analysis – whether it's praise for a company's ESG efforts or revelations of inconvenient truths – has been done manually. The process was therefore error prone, slow and very costly.
Fighting information asymmetries
So it's no surprise that many financial services firms are exploring ways to use artificial intelligence (AI) technologies to improve ESG ratings. AI technologies can excel in three areas:
- News monitoring. Algorithmic-based systems can easily mine large amounts of unstructured data through automation. They can be configured to monitor events, effectively scanning the worldwide web, as well as curate and gather valuable company information from a variety of sources, including social media, daily local news and newly available reports. Only automation can ensure fast and timely collection of relevant data.
- Improved reporting. An AI engine trained with appropriate and user-defined ESG criteria can quickly analyze and convert the quantitative and qualitative data collected into structured and actionable information. A well-designed system can use data to compile ESG reporting frameworks quickly and automatically. You can also identify any information gaps. These activities help improve reporting, making the intelligence gathered much easier for analysts to digest.
- Sentiment analysis. Natural language processing technologies – a branch of AI – are able to analyze the contextual, semantic and sentiment factors embedded in the collected datasets. As a result, it is now possible to discern the tone of the information provided and classify it, for example, into 'positive', 'negative' or 'necessary action'. Analytical algorithms could be trained to examine a certain type of conversation and identify its tone by comparing the words used against an existing reference set of information, such as 'child labor' or 'modern slavery'. Also important is the concept of “expert in the cycle”: AI-driven systems in which humans remain in full control and their supervision is active and involved. In this case, ESG experts and analysts are able to train the machines by providing their feedback on a story. For example, the questions experts answer may include “is this news relevant?” and “does it belong to one of the following categories: 'environmental', 'social', 'governance' or 'controversial news'?”. Over time, the AI engine will become better and better at understanding the sentiment of new information received, as can be seen in the figure, which shows the process in detail.
A better world
Using AI for ESG purposes should produce one “win-win-win” scenario in terms of profit, planet and people. However, we are only at the beginning of using AI technologies to drive ESG improvements. As a result, it's probably too early to tell whether AI can really help move the ESG agenda forward. However, the odds are good, as history is filled with examples of how technologies have helped us achieve social goals and create a better society.
Sadly, many past sustainability efforts have ended up as “fads” or slogans claiming to do good. But the problem seems to come from us and not from the technologies we have used. By empowering investors, especially those with an ESG focus, it is possible that ESG will become a common and respected corporate practice. Using AI in the right way can contribute to our effort to do well and do good at the same time.
