An Intro to Artificial Intelligence in ESG Investing

By GLOBUS Correspondent, Diogo Ribeiro Dos Santos

ESG investing continues to reach new highs, with assets under management increasing exponentially and a large percentage of all worldwide investments being allocated towards it. In my other articles, I explore how ESG investing works and its benefits, as well as the controversies and issues surrounding it. In this article, I aim to explore the role and presence of artificial intelligence (AI) in aiding us in making smarter ESG investments, alongside its current limitations and drawbacks.

One of the current limitations of current ESG investing is the lack of data, and lack of uniformity of existing data. AI allows us to tackle the latter issue. It allows investors to collect and analyse more information than ever before when accounting for environmental, social, and governance risks and opportunities. It can unearth key data for investors seeking sustainable investments who are surrounded by a plethora of unreliable data, by utilising machine learning algorithms to automate complex tasks at incredible speeds.

As of now, not only is there a lack of uniformity in ESG reporting standards, but there’s a lack of ESG reporting at all. AI can help us interpret alternative data such as textual data, to identify companies’ controversies and important ESG news, and satellite data to verify companies’ carbon emissions, or analyse their impact on ecosystems. However, company disclosures can be subject to manipulation especially considering AI algorithms, and a lack of historical data in some instances, which might lead to biases and representative issues. Thus, it’s essential a comprehensive investment process avoids placing too much confidence in a single ESG metric.

In 2012, a small team of entrepreneurs, data scientists, and engineers founded Rho AI, a data science firm, attempting to leverage artificial intelligence to encourage environmental, social and governance investments to limit the impact of climate change. Today, their products are used in a range of industries, including healthcare, finance, sports, waste, water, climate change, and energy. They started developing custom software for professional motorsports to predict an optimal race strategy. Now, they utilize an open-source artificial intelligence tool integrating automated web scraping technology and machine learning with natural language processing to enable investors to evaluate the climate impact of companies and to use this evaluation as a basis for making investments in companies.

In their pursuit to develop this vision, they found 3 large barriers. Unlike traditional financial data that were structured and quantitative, ESG data tended to be ‘unstructured, qualitative, scattered, and incomplete’ greatly complicating the process of interpreting available information. Evaluation of ESG data required more robust tools to ensure that they were used appropriately, and they found Investors were not using new tools of data collection and analytics effectively. This is understandable considering the complexity of constantly needing to learn to interpret new data in new forms.

Additionally, they were faced with the question of AI and ethics. As Rahwan et al. (2019) stated “Machines powered by artificial intelligence increasingly mediate our social, cultural, economic and political interactions”, raising the question of that what extent should we allow AI to mediate our interactions to such a level?

The Rho AI team found 4 principal concerns. Although the code could be simple, the results could be complex, resulting in ‘black boxes’ – the paradox where even if the code is simple, engineers/coders cannot understand how AI reached its conclusions due to a lack of process transparency. Moreover, the code may not be simple and in fact extremely difficult to interpret further exacerbating the black box problem. Imperfections in data could impact on the results produced, going back to how algorithms, models, and formulas are only as good as the inputs they receive. As noted earlier, data is prone to manipulation and often encompasses willing and unwilling errors. Lastly, much of the source code and training data were proprietary, diminishing the ability to develop fair, effective AI.

Adopting AI for any purpose can pose a significant environmental impact. The process for creating and training AI algorithms requires large amounts of computing power, which in turn consumes large amounts of electrical energy. Additionally, there are grave concerns on the ethics surrounding AI use. In 2019, Google created an ethics advisory board to guide its research into and use of AI but had to quickly disband the board due to controversy over some of the board members. Also in 2019, the European Commission introduced a set of Ethics Guidelines for Trustworthy Artificial Intelligence due to concerns regarding fundamental freedom and human rights such as privacy, personal data protection, and algorithms focused on profiling people.

Nevertheless, AI has definitively helped numerous industries and businesses scale and become more efficient. Within investments, it has aided analysts in interpreting multitudes of complex data at much faster rates. And in ESG investing, it’s starting to be used and tested although many limitations remain and it may still be some time until we see standardised, practical applications being put into place.

Header image by Possessed Photography via Unsplash

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Website Powered by

Up ↑

%d bloggers like this: