It is understandable that some people should be sceptical about artificial intelligence (AI) but there is no shortage of ways companies can use technologies not just to do good but also to do well, like using AI to help reach circular economy goals. Yet the progress for technology-driven pursuit of sustainability remains slow.
One reason for such sluggishness is that there remains a lack of economic incentive for companies – and their investors – to make huge investments concerning social and environmental gains and benefits. This also provides an explanation as to why, despite years of discussions on the importance of the so-called “triple bottom line” – the need to care for not just profit but also people and the planet –, it has hardly become a mainstream practice among today’s businesses. At the same time, there is increasing interest in investing in socially and environmentally responsible companies.
Looking at it from this vantage point, two lessons are clear:
- The first is that unless investors are getting the satisfactory return, it will be difficult to get businesses to orientate themselves towards goals related to people and the planet.
- The second is that investors need to have timely and accurate information to make informed decisions.
In this respect, AI presents a welcome and potentially extremely beneficial tool to help the latest idea in sustainability: environmental, social and corporate governance (ESG).
Sustainable investments reached $31 trillion in 2018
Investments in ESG have fast become a socially important asset class. One study points out that sustainable investments reached $31 trillion in 2018, up by 34 percent from 2016. The result: we need to know if and how different companies are acting in accordance to ESG. Simultaneously, as revealed in a recent survey, it is now becoming evident to managers that ESG is an essential ingredient to the long-term survival of their businesses. Not surprisingly, with the raised interest in the subject, a number of companies such as MSCI, Bloomberg and Sustainalytics have proliferated in recent years to offer ESG-based rating and index services to guide investors.
Boohooing ESG ratings
Yet these ratings can be heavily flawed. Take for example, the ultra-fast-fashion retail phenomenon Boohoo. The retailer received well-above ESG ratings before being accused of modern slavery for underpaying their workers in Leicester (England!) and not providing them with proper protective equipment against Covid-19. Another example is Wirecard, a german payment processor and financial services provider. The company went bankrupt in June 2020, but news of its dodgy and questionable business practices had been uncovered as early as 2015. Yet throughout all this time, Wirecard received median-grade ratings from a number of ESG ratings agencies.
How could the rating companies have gotten it so wrong? The answer: information asymmetry. It appears that all parties involved face different challenges with obtaining and processing information. To begin with, rating producers and indices deploy their own proprietary methodologies and data to analyse companies. The result of them using different ESG definitions, measurements and weightings for different indicators often lead to conclusions and verdicts that can be distinctly different from one index to another. A recent study has found that in a dataset of five ESG rating agencies, correlations between scores on 823 companies were on average only 0.61. Furthermore, they rely heavily on information provided by the companies being rated. In this case, the rating companies are essentially giving the latter a free pass to pick and choose the data to be used.
This, in turn, poses problems for the investors. First, without standardisation across ratings, it is difficult for investors to compare across the indices created by different providers. Another key problem the investors face is that they have got little choice but to rely on the rating and index producers to capture the latest information and news and to incorporate them into their ratings. The lack of timely and accurate information makes it difficult to make the right decisions. Take the example of Tesla. Is this electric car company really an ESG company considering their batteries depend on nickel, the extraction of which comes at an environmental and health cost? What about Tesla’s recent purchase of some $1.5 billion worth of bitcoins, which require a great deal of energy to process?
For the rated companies, even though they can select the information to be submitted, they often suffer from other problems. For example, how can they improve their own ratings if different raters use different methodologies? Also, how can they ensure the investors will get hold of the right information and interpret it correctly?
In short, the problems emerge from a lack of clarity, consistency, and transparency of ESG ratings as well as information asymmetry and shortage.
AI as a tech for good
The key to solving these problems is to regularly and rapidly collect qualitative information to supplement the publicly available data. Up-to-date qualitative data has the ability to not only help investors and rating producers to be much better informed, but such data can also be used to set up key inputs that could be used as the basis to form common minimum standards.
Until now, gathering and gleaning insights from social media, daily local news, and freshly available reports has been carried out manually, which is error-prone, slow and very costly.
Algorithm-driven systems are a potential game-changer in pushing the ESG agenda as they can easily and effectively crawl the worldwide web and scrape unstructured data on companies from a range of sources. They can then swiftly parse and convert the excavated data into usable structured ones. This, in turn, allows for curated output that is valuable for all parties involved, thereby drastically mitigating the information asymmetry problem. In addition, using natural language processing technologies to perform analyses that capture sentimental, contextual and semantic factors embedded in the collected data, it will be possible to discern the tone of the information provided. Such analytical algorithms could be trained to go through a certain type of conversation and identify the tone by comparing the words used to a reference set of existing information.
Something for everyone
This should produce a “win-win-win” situation:
- For the rating and index providers, real-time signals often represent early warnings and timely indicators, giving them much clearer pictures of the companies in their indices. They can also broaden the scope of analysis using AI-derived information to complement their current quantitative data and methods.
- For the rated companies, they can be assured that their ESG-compliant activities will probably be captured and integrated in indices and ratings.
- Finally, for investors, they will have better quality ESG data quicker, which in turn can help them with their investment strategies.
It must be noted that we are only at the very beginning of using AI technologies towards ESG improvements. And the results of what this would succeed is yet to be seen. However, history is filled with examples of how technologies have helped attain social goals and create a better society. It is very much hoped that through the empowerment of investors, particularly those who are ESG-orientated, it will become possible for ESG to be a well-respected and common business practice instead of just another fad or slogan advocating sustainability. Using AI in the right way can no doubt help with our endeavour to raise ethical standards.
This article is also published by The Choice. Energy Voices is a democratic space presenting the thoughts and opinions of leading Energy & Sustainability writers, their opinions do not necessarily represent those of illuminem.
Terence Tse is a professor at Hult International Business School and executive director of Nexus FrontierTech. Global speaker, and has written multiple published articles and books, including The AI Republic: Building the Nexus Between Humans and Intelligent Automation.