A question of trust: How AI is addressing greenwashing concerns
By helping to allay rising greenwashing concerns, technology-driven ESG assessment could play a vital role in bridging the funding gap for climate and sustainability initiatives.
Greenwashing – the overstating of environmental claims – is one of the biggest problems facing climate-focused investors. According to the 2023 Google Cloud Sustainability Survey, over 70% of executives believe that most organizations in their industry would be found guilty of greenwashing if investigated thoroughly. Perhaps more surprisingly, almost 60% admitted to exaggerating their own sustainability activities.
Naturally, greenwashing risks faced by companies – including reputational, litigation and regulatory fallout – are on the rise around the world, and especially in Europe and North America (see Figure 1).
Greenwashing poses particular challenges for investment professionals. In addition to having to navigate greenwashing concerns around their own products, asset managers are also exposed to litigation and regulatory risk related to greenwashing by companies they invest in (see Figure 2).
Accusations of greenwashing levelled against banks have become far more frequent between 2021 and 2024. And US sustainable funds saw a record USD8.8 billion in net withdrawals in the first quarter of 2024, attributed to a variety of factors including greenwashing concerns.
Misdirecting investment
More broadly, greenwashing erodes investor trust and results in misallocated assets. Through exaggerated green claims, companies could attract funds they are not deserving of. And as greenwashing erodes the general perception of sustainability endeavors, it can lead to deserving practices being underweighted, said Monika Freyman, Vice President, Sustainable Investing at Addenda Capital, on a recent CFA Institute webinar: Building Sustainability Credentials – Can AI End Greenwashing?
Compounding the problem, there is a lack of standardization in companies’ environmental, social and governance (ESG) information, with ratings from third-party agencies often at odds with each other. A 2022 Dow Jones survey of 200 financial leaders revealed that 58% of respondents feel there is a need for greater transparency around how ESG ratings are scored.
This lack of standardization can also lead to unintentional greenwashing, said another speaker on the webinar, Ángel Agudo, VP of Product at Clarity AI, a sustainability technology platform that uses machine learning and big data to deliver ESG insights to investors and organizations.
In light of these concerns, investors are increasingly looking to form their own views on companies’ ESG performance.
“When you are trying to identify and avoid greenwashing, you want to look at a lot of relevant data very quickly. Technology makes that a lot easier,” said David von Eiff, Director, Global Industry Standards at CFA Institute.
Several firms, including Clarity AI, offer tools that enable financial institutions to apply artificial intelligence (AI) to ESG data from a wide range of sources to form more accurate, objective and consistent assessments of companies’ climate and sustainability commitments and performance.
Pledges vs. reality
Many of these solutions seek to tackle greenwashing by comparing ESG pledges and claims to actual actions and progress. They do this by incorporating a combination of structured and unstructured data – the latter is made available using a subset of AI known as natural language processing (NLP), which enables computers to understand text and speech.
According to Agudo, a company’s ESG performance can be assessed using data from three main sources:
- Company disclosures, from its corporate website, sustainability reports or filings. “As of today, we need to trust what they are reporting is factual,” he said.
- Contextual company information, such as sales numbers, which are useful in assessing the reported ESG data.
- Information from a wider context, such as social media, news, independent research, and reports from non-governmental organizations. “This can serve as a sort of audit of the company’s own information, and help analyze how the company is doing from a broader perspective,” said Agudo.
Where there are gaps in information, estimates can be used, but it must be made clear that those are estimates, stressed Agudo.
Indeed, transparency and traceability of all data is paramount, with links provided to original sources so users can make their own judgements about the relevance and veracity of those data points. After all, exercising human judgement is vital to avoid being led astray by mistakes, gaps, inconsistencies and biases in data.
Critically, these solutions also tend to offer a more granular view of companies’ and funds’ ESG performance than the ESG scores offered by ratings agencies, which Agudo likens to going to the doctor and only being told that your health is “70 out of 100.” That, he says, is not as helpful as a granular view of each of your body’s systems. For instance, your heart could fine, but your lungs might need attention.
“In ESG there are a lot of dimensions we need to take into consideration. The environmental factors alone are very complex,” said Agudo. “That complexity cannot be reduced to one single number, especially if you want it to be actionable.”
No silver bullet
Still, it is important to be aware of the limitations of AI-driven ESG assessments. Gary Gensler, Chairman of the U.S. Securities and Exchange Commission, cautioned businesses against making exaggerated claims about AI’s ability to solve problems, which he dubbed “AI washing.”
AI-based ESG assessment solutions are also beset by problems commonly associated with AI, including a lack of explainability and contextual understanding, said Agudo. But he believes the technology will get better over time, “delivering better insights, explainability and reporting.”
One of the key drivers of that improvement could be the incorporation of unstructured data from many more sources, including technology that is being increasingly deployed to drive real-world emissions reductions. For example, ESG assessments might incorporate data from AI-powered satellite data platforms that are used by companies to improve the tracking, tracing and reduction of their emissions. Or data from internet of things (IoT) sensors used to improve energy efficiency.
Integrating data from sources like these could give companies and investors a much clearer picture of climate and sustainability performance, providing much-needed information to support their decisions. For climate-focused investors, advanced technology also has the potential to improve how they assess the credibility and feasibility of transition targets.
As better data and analysis leads to more accurate assessments of ESG performance and allays greenwashing concerns, more investment is likely to flow to bona fide climate and sustainability initiatives. This could help narrow the enormous unmet financing needs for climate action and the United Nations’ Sustainable Development Goals, estimated at around USD4 trillion annually.
You may also be interested in
14 Nov 2024
How the finance industry influences ESG impact
.jpg?crop=yes&w=768&h=512)
12 Nov 2024
Why climate change investing is here to stay
Want to learn more about ESG Investing?
Choose your own learning path through our sustainability certificates: the foundational Certificate in ESG Investing and the advanced Climate Risk, Valuation, and Investing Certificate.
