This paper is part of the Fide Foundation’s GET-2 ESG Think-Tank final report from the 2024 Oxford Congress, titled “Driving Change: Exploring Opportunities and Challenges in Accelerating Sustainable Finance.”
Key Findings
Artificial Intelligence (AI) brings multiple benefits to sustainable investing, in different forms, and through different use cases with different levels of maturity and complexity (e.g. data collection and processing, predictive analysis, portfolio optimization, data visualization, image recognition, scenario analysis, and enhanced risk analysis).
The availability of quality data is of the utmost importance. Not only can AI collect reported data in a more scalable way -reducing human errors- but it can also source alternative data to validate information reported by companies and build estimations where needed, through sophisticated models, trained and supervised by humans.
Risks are also to be considered when deploying AI, and more and more, investors are considering responsible approaches towards it.
The private sector has gone through considerable investments and developments in the field, while the public sector, led by some key institutions, is also contributing with interesting resources and knowledge.
Abstract
The integration of AI in sustainable investing is transforming how investors access, analyse, and use sustainability data. It brings many benefits, such as a higher level of sophistication for data analysis and predictive tools. Though, it also comes with certain challenges such as data quality, models’ transparency, or interoperability questions.
Nowadays, adopting AI for sustainable investing can only be successful in combination with human expertise, to ensure the tools are really fit for purpose, ultimately supporting investors in making more informed decisions without adding unnecessary complexity to their current processes.
Due to its enormous potential, both the private and public sectors are increasing their efforts to leverage the value of AI in sustainability and sustainable investment.
Key Words
Artificial intelligence, sustainable investment, data quality, data availability, human intervention (human in the loop).
Content
Introduction
In this panel discussion, we looked at how AI is reshaping sustainable investing by overcoming traditionally unsolved challenges associated with ESG or broader sustainability data and insights. With more companies reporting, mainly driven by regulations across the globe, AI’s capacity to scale data collection efforts and ensure reliability is critical. Especially, when discrepancies across data providers and incomplete corporate reporting remain significant issues.
In any case, the capacities of AI in this field go far beyond data collection and ESG reporting, and include a wide range of possibilities, such as converting unstructured data into useful insights, conversing with your portfolio, and beyond, as we will see in the next section.
Sustainable investment and AI: benefits and challenges
AI is becoming a differentiating tool when it comes to sustainable investment1, and it would not be possible to approach the AI topic without also addressing the relevance of data.
- Data challenges for sustainable investing:
Investors still face challenges related to sustainability data used for decision-making and reporting. Below we highlight some of the main ones, directly connected with the role of AI:
- Data availability and accessibility tend to be a widespread problem. In some cases, data is inaccessible (e.g. proprietary data, data from the public administration that has not been shared as open data or it is limited either in its scope or quality). And that is a relevant question, especially considering that investors need to look beyond reported data (e.g. alternative data sources) to get a fair representation of a company’s performance.
- Data quality (gaps, errors, issues to integrate and standardize different data sources)2 is a common issue too. And in this sense, there are different dimensions that must be considered when talking about data quality (e.g. timeliness, completeness, uniqueness, consistency, validity, accuracy).
- Opaqueness: regarding the relevance of data transparency, in 2022, Forrester, described how opaque ESG ratings were no longer satisfying investors, regulators, and corporates. It also highlighted the importance of: (i) data granularity that serves specific investments and for a better risk management, (ii) data interoperability and standardization, and (iii) disclosure capabilities for reporting purposes3.
- The role of AI to accelerate sustainable investing:
“AI technologies play a crucial role in ESG investing by enabling investors to analyse ESG-related data, enhance decision-making processes, and align investment strategies with sustainability goals.”4
Amongst the benefits it brings, we find that using AI for sustainable investment:
- can be leveraged for multiple purposes: predictive analysis, portfolio optimization, data visualization, image recognition, scenario analysis, and enhanced risk analysis, amongst others.
- makes it possible to combine and analyse enormous sets of data (including unstructured data from many data sources), to identify parameters and hidden dynamics, trends, and patterns, and to perform a more precise and in-depth analysis. When reported data is not available, AI can help source alternative data and estimations to fill gaps.
- when connected to real-time information sources, it allows, for example, to adjust investment portfolios considering changing ESG conditions (dynamic portfolio optimization).
According to the IIF5 “AI’s capabilities to process massive sets of data through increasingly sophisticated methods (e.g. machine learning, large language models for generative AI, and natural language processing) can identify patterns and make predictions that assist in assessing the impact of sustainable investments. This helps financial firms make more informed decisions and better gauge the risk-return profile of sustainable investments—and ward off accusations of greenwashing and other reputational concerns.”
- can increase the efficiency and effectiveness of ESG and compliance teams in filing regulatory reports, shifting their time to more added-value tasks such as product innovation or client service.
- Challenges to implementing AI in sustainable investing
But the use of AI in sustainable investing also comes with some challenges, many of which are common to other AI spaces:
- In certain cases, AI models can have complex explainability and interpretability. And that can affect risk management in the use of AI models.
- Related to transparency and other topics that touch on “responsible AI,” the World Economic Forum has analysed how different stakeholders can benefit from the use of responsible AI governance and techniques (please see the chart below).
- The quality of the data used in training AI models is crucial to ensure the quality of the output, and avoid what is known as the “garbage in, garbage out” concept, which refers to poor-quality input producing faulty output.
- Regulatory fragmentation (both for AI, data, and ESG) is a legal but also an operational question that affects AI models.
- The environmental cost of AI must be accounted for. Firms need a clear understanding of when AI is the right strategy to achieve a goal to avoid inefficiencies and waste. Additionally, companies can implement mitigation measures to minimize AI’s impact on the environment.
Source: Responsible AI Playbook for Investors6 (2024)
How AI is helping to expand the impact and reach of sustainable investment
There are many ways in which AI is boosting ESG investment. For instance, some AI tools are used to identify and mitigate ESG-related risks7. Some of them leverage the benefits of predictive analysis through scenarios (e.g. climate and others) and are used to forecast future trends. Others are being used for portfolio optimization purposes, considering their performance and financial returns.
A recent analysis by Clarity AI looked into impact funds8 to determine the alignment between the focus of investments and the areas that require the most attention, using the United Nations Sustainable Development Goals (SDGs). The research found that companies in SDG funds sell less than 1% in the countries that need the most. AI can support advanced analytics, for investors to access a more granular, exhaustive analysis to improve capital allocation.
Sentiment analysis is a very potent tool too. E.g. analysing news, social media, and other sources (including investors’ interactions) to gauge public sentiment towards companies’ ESG practices.
Generative AI can also help bridge the gaps in ESG skills by supporting analysts in navigating methodologies, interpreting results, and deciding a course of action. As mentioned previously, it can also reduce the time spent in regulatory reporting and disclosures, reducing human errors and increasing team efficiency.
Lastly, other examples include robo-advisors providing investment advice based on ESG scores and data, and ESG algorithmic trading.
Involvement of public institutions: recent use cases
Together with many valuable and scalable solutions from the private sector, it’s worth mentioning some initiatives from the public sector that leverage on AI and the value of data: Malena from World Bank and Project Gaia, Project Viridis and NGFS Data Directory 2.0 from the Bank for International Settlements (BIS) and others.
Conclusion and proposals
The use of AI for ESG investment is not new, but its scope and capabilities keep on growing, and its potential in terms of portfolio optimization, risk management, and data analytics is unparalleled. While AI can help solve traditionally unsolved challenges around data, data availability and quality are in turn essential to improve and materialize the impact of AI on sustainable investment.
What technology shall be used for each use case is a technical but essential question. In many cases, predictive tools capable of detecting patterns, performing classification, and prediction are and will continue to be used, while still, it is to be seen what the impact of relatively recent technologies such as generative AI (GenAI)9 will be for the sustainable investment field.
Beyond private sector initiatives, public sector initiatives are also arising to materialize the value of this technology for investors, supervisors, and the financial sector in general, and ultimately all stakeholders.
Links to recommended readings or specific bibliography on the content of the panel.
- World Economic Forum and CPP Investments Insights Institute (2024). Responsible AI Playbook for Investors Whitepaper.
- Credible climate transition plans: Insights from an AI-driven analysis of corporate disclosures. Nico Fettes, Climate Research Director, Clarity AI. June 2024.
- The next frontier in investment management: GenAI for smarter, more efficient portfolios. Clarity AI. August 2024
- Enabling efficiency in sustainable fund regulatory reporting. Clarity AI. August 2024.
- The missing GHG emissions: How satellite data can quantify the real climate risk of oil & gas companies. Patricia Pina, Head of Product Research & Innovation, Clarity AI, and Renato Coelho, former Director of Research, Clarity AI. July 2024.
- How advanced technology can increase data reliability. Research & Data Science teams, Clarity AI. November 2022.
- Abdalmuttaleb M. A. Musleh Al-Sartawi, Khaled Hussainey & Anjum Razzaque. (2022) The role of artificial intelligence in sustainable finance. Journal of Sustainable Finance & Investment 0:0, pages 1-6.
- Whelan, T. et al. NYU Stern and Rockefeller Asset Management (2021). ESG and Financial Performance: Uncovering the Relationship by Aggregating Evidence from 1,000 Plus Studies Published between 2015 – 2020
- Tiftik, E. et al., Institute of International Finance. (2020) “IIF Green Weekly Insight: ESG Funds Deliver”
- Fu T and Li J (2023), An empirical analysis of the impact of ESG on financial performance: the moderating role of digital transformation. Front. Environ. Sci. 11:1256052. doi: 10.3389/fenvs.2023.1256052
- Institute of International Finance (IIF). (2024). Green Weekly Insight: Leveraging AI for a Better, Faster Net Zero Transition https://www.iif.com/publications/id/5807
- Cruces Rufo, R. (2020). Data: unfinished business: Leading business-driven data management and governance transformation.
- KPMG (2024) The Use of AI in Sustainable Finance (kpmg.com)
Participants
Gloria Sánchez Soriano, Senior Advisor at Institute of International Finance (IIF).
Lorenzo Saa, Chief Sustainability Officer, Clarity AI
Moderator: Ana Rivero, Operating Partner ESG, Alantra. Member of the Scientific Committee Oxford 2024.
Furthermore, this document is signed in a personal capacity and does not represent the official position of the institutions or entities to which the authors may belong.
Oxford/24 Final Report:
This paper is part of the Fide Foundation’s GET-2 ESG Think-Tank final report from the 2024 Oxford Congress, titled “Driving Change: Exploring Opportunities and Challenges in Accelerating Sustainable Finance.” Held at Jesus College, Oxford on September 18th, 19th, and 20th, 2024, the Congress brought together world leaders in finance, regulation, and sustainability. This comprehensive report consolidates key insights from the event, offering strategic recommendations to financial institutions and regulators on transitioning to a low-carbon economy and reaching net-zero greenhouse gas emissions by 2050.
The full report can be found at: https://bit.ly/oxf24-report
