Banks use AI for EU Taxonomy alignment
Banks are using artificial intelligence for ESG disclosure and reporting and assessing EU Taxonomy alignment is emerging as one of the strongest use cases to date as it underpins EU regulation.
Requirements are now live for the EU Taxonomy and European Banking Authority's Green Asset Ratio, which requires banks to show the share of EU Taxonomy-aligned assets in selected financial assets in their Pillar 3 reports this year.
Banks are also preparing for the Corporate Sustainability Reporting Directive, which requires companies to start reporting as required for the 2024 fiscal year, with the first report to be published in 2025.
"Banks can use artificial intelligence, and specifically generative AI, for a mix of EU Taxonomy alignment, Pillar 3 and CSRD reporting requirements," said Olivier Vinciguerra, CEO and co-founder of technology platform Amalthea.
European banks face lengthy and complex processes to monitor and report on the performance and alignment of their lending activity, especially in areas such as asset finance and project finance, and financial institutions are either building their own systems or buying tools to help them manage the process.
Top-tier banks are generally choosing to create their own climate data platforms, while second-tier banks are choosing to buy tools from a range of data providers and fintech companies to manage the already overwhelming amount of data.
CSRD reporting is starting with large companies and parent companies of large groups with more than 500 employees, and banks are assessing all of their loans to those firms to monitor their EU Taxonomy compliance and the relative greenness of their banking book.
This is highlighting relatively low levels of Taxonomy alignment in European banks' loan portfolios, according to Amalthea, which also helps to evaluate bonds in the context of the incoming EU Green Bond Standard that will be implemented on December 20.
"Banks are reporting their EU Taxonomy alignment for their largest companies. It's not high so far, maybe 8%–10%," Vinciguerra said.
The extension of CSRD to midcaps, SMEs and non-EU companies from 2025 to 2029 will increase the workload for banks and make the process nearly impossible to manage on a manual basis.
"At some point banks will be slowed by the volume of loans that they need to run for CSRD and they will need artificial intelligence to accelerate their processes and minimise their costs," Vinciguerra said.
Banks with their own net-zero targets in place or that are members of the Net Zero Banking Alliance also have to monitor their lending books closely as they seek to transition their clients and reduce their exposure to high-emitting sectors and potential stranded assets.
"Banks need a lot of data to understand how their banking book is going to evolve and how the different companies that they are financing are going to reduce their emissions over time," Vinciguerra said.
Amalthea is one of a range of data providers that use artificial intelligence to assess EU Taxonomy alignment, which also includes larger firms such as Sustainalytics and MSCI. Standard & Poor's offers the service as part of its second-party opinion offering.
The firm uses cloud-based retrieval-augmented generation and large language model technology and has integrated Mistral AI, the French version of ChatGPT, to assess the Taxonomy alignment of loans and potentially reduce operating costs.