Financial institutions are having their collars felt by regulators on greenwashing. Last year, there was a 246% increase in fines related to environmental, social and governance issues, according to data collated by the regtech Fenergo.
Environmental disinformation, as many regard it, is a front-and-centre issue, particularly in the US, where we have already seen the Securities and Exchange Commission imposing multimillion dollar fines on FIs for greenwashing. While FIs are busy scrambling to get their act together, there is still no harmonised and global legal definition of greenwashing.
Crucially, this continually amorphous concept varies by product, service, regulator and jurisdiction.
The UK's Financial Conduct Authority interprets it as "marketing that portrays an organisation's products, activities or policies as producing positive environmental outcomes when this is not the case".
The definition from the International Organization of Securities Commissions, the global standard-setter for securities markets regulators, is subtly different.
It labels the attempt to capitalise on the growing demand for environmentally sound products as "the practice of misrepresenting the sustainability-related features of investment products".
This lack of definitional precision and alignment risks the sustainability claims of FIs being treated as greenwashing by some, but not all, regulators.
Despite this variance in definitions, a common theme can be settled on: greenwashing is about misrepresentation, misstatement and false or misleading practices in relation to ESG credentials.
Addressing this risk is a growing concern for FIs, because it is becoming an increasing source of litigation and regulatory scrutiny.
The FCA, for example, is introducing an 'anti-greenwashing' rule in 2024, which is applicable to all FCA-authorised firms.
It will demand that "all claims made about the sustainability characteristics of a product or service must be clear, fair and not misleading".
The EU has also introduced legislation aimed at tackling greenwashing that applies to FIs, such as the Sustainable Finance Disclosure Regulation and the Taxonomy Regulation.
Although the US is somewhat behind the EU in its pursuit of tackling greenwashing from a regulatory perspective, the SEC has proposed climate disclosures rules that, if adopted as proposed, would regulate the climate-related reporting of US public companies.
The SEC has also launched a Climate and ESG Task Force within its Division of Enforcement to develop initiatives aimed at proactively identifying ESG-related misconduct.
By introducing anti-greenwashing measures, regulators are attempting to standardise ESGrelated investment product labels and establish classification systems to determine whether economic activities can be described using terms such as 'environmentally sustainable' and 'green'.
The problem is that what is deemed as environmentally sustainable or green is open to interpretation across different jurisdictions.
This developing regulatory landscape, and the associated litigious and reputational risks associated with being accused of greenwashing, are making ESG reporting an increasingly worrying issue for FIs.
Given the breadth and detail of ESG reporting, it has become difficult for FIs to identify greenwashing without some use of technology. With this in mind, adopting generative artificial intelligence may power firms with greater speed and a higher degree of scalability than current technology.
What is GenAI for greenwashing?
While large language models can have a general purpose, they can also be trained on specific data which focuses the models on a particular practice area, enabling responses which can include commenting on adherence to regulatory obligations.
In a greenwashing context, some LLMs have been trained on qualitative publicly available data and regulatory guidance regarding a range of ESG issues.
Using this data, the LLMs are trained to identify potential instances of greenwashing and flag these in the response to the user.
One simple example is where a sustainability claim has been made by a FI (possibly on its website, in its pre-contractual documentation or in its periodic reporting, as required under the Sustainable Finance Disclosure Regulation), but the actual practice and evidence differs from the reported assertion.
Here, an LLM can be trained to identify the difference between reported assertion and fact to flag to the end user for further investigation.
Why use GenAI?
Whereas human review of a FI's ESG-related reports may take days, an LLM can reason over documents in minutes.
While some level of human expertise may be required to review an LLM's response, adopting GenAI to complete early-stage review has the potential to save significant cost and time.
We tested this assertion by prompting a general-purpose GenAI model to review a publiclyavailable sustainability report of a large multinational company and output any potential instances of greenwashing.
In less than two minutes, the model outputted eight reasonable potential examples of greenwashing.
Although the model did not cross-reference external evidence or regulation, its ability to 'memorise' the report and identify contradictory statements was shockingly effective.
As well as the potential benefits to ESG investors in making investment decisions, the advantages of adopting this tech will be obvious to firms engaged in climate litigation, who can leverage the speed of GenAI to assess potential greenwashing over a much larger number of third-party chains.
Risks of using GenAI to identify greenwashing
While GenAI can streamline an initial review for greenwashing, care (and cost) must be taken to ensure the accuracy of an LLM's output.
The 'robustness' and accuracy of output is a concern. GenAI has exhibited the capacity to output information that is incorrect (while still sounding plausible) and then subsequently defend the accuracy of the output.
This phenomenon is known as 'hallucination'. In relation to greenwashing, an example may be where underlying data is cited in an ESG report but cannot be verified without real-world environmental assessment.
If asked to comment on the veracity of this data, a model may hallucinate a response which purports to rely on real-world data which in fact does not exist.
Proceeding with caution
As greenwashing-specific LLMs develop, risks concerning output accuracy and the need for human verification should decrease. ESG-related datasets should be carefully selected to ensure that training mitigates these concerns. FIs should carefully weigh up their ethical and governance policies on using AI.
Increasingly burdensome ESG reporting requirements and the need for a scalable solution naturally lead us to GenAI, which is capable of consolidating information and identifying potential greenwashing at superhuman speeds.
While the solution is not infallible, the technology will continue to improve and the risks associated with GenAI may become less concerning to FIs.
Originally published by Financial Times - Banking Risk & Regulation.
Originally published by 04 March, 2024
Visit us at mayerbrown.com
Mayer Brown is a global services provider comprising associated legal practices that are separate entities, including Mayer Brown LLP (Illinois, USA), Mayer Brown International LLP (England & Wales), Mayer Brown (a Hong Kong partnership) and Tauil & Chequer Advogados (a Brazilian law partnership) and non-legal service providers, which provide consultancy services (collectively, the "Mayer Brown Practices"). The Mayer Brown Practices are established in various jurisdictions and may be a legal person or a partnership. PK Wong & Nair LLC ("PKWN") is the constituent Singapore law practice of our licensed joint law venture in Singapore, Mayer Brown PK Wong & Nair Pte. Ltd. Details of the individual Mayer Brown Practices and PKWN can be found in the Legal Notices section of our website. "Mayer Brown" and the Mayer Brown logo are the trademarks of Mayer Brown.
© Copyright 2024. The Mayer Brown Practices. All rights reserved.
This Mayer Brown article provides information and comments on legal issues and developments of interest. The foregoing is not a comprehensive treatment of the subject matter covered and is not intended to provide legal advice. Readers should seek specific legal advice before taking any action with respect to the matters discussed herein.
