Project Gaia: enabling climate risk analysis

  • An artificial intelligence (AI) application developed by the Bank for International Settlements (BIS) and its project partners, the Deutsche Bundesbank and the European Central Bank, specializes in extracting and structuring data from text-based sources, such as reports and publications.
  • Project Gaia created a proof of concept and tested it in financial climate risk analysis use cases, enabling analysis of climate-related risks in the financial system.
  • The proof of concept used a flexible architecture and variety of large language models (LLMs) to automatically and robustly extract climate-related indicators from publicly available corporate reports.

Project Gaia, an initiative of the BIS Innovation Hub, the Deutsche Bundesbank and the European Central Bank, explores how AI-driven text extraction can provide high-quality, accessible data at scale for a wide range of applications in money and finance. One focal point has been climate-related risks, where a lack of global reporting standards makes comparison difficult.

During the first phase, in collaboration with the Bank of Spain, Gaia broke new ground by integrating LLMs into an application for extracting data on climate-related financial risks. The flexible design may serve as a model for AI-enabled applications in a broader range of use cases for central banks and the financial sector.

Central banks, supervisors and financial institutions need high-quality and accessible data to model the financial risks posed by climate change. A lack of global reporting standards makes this difficult. The project report describes how Gaia was able to overcome differences in definitions and disclosure frameworks across jurisdictions to offer much-needed transparency and make it easier to compare information on climate-related financial risks.

The experiment was coordinated by the BIS Innovation Hub Eurosystem Centre and aimed to help analysts and supervisors search corporate climate-related disclosures and extract data quickly and efficiently on indicators such as total emissions, green bond issuance and voluntary net-zero commitments. Using AI and, in particular, LLMs, Gaia delivered a proof of concept demonstrating it is possible to automate the task of identifying such indicators across a large set of reports, significantly reducing manual effort in climate assessments.

The second phase of Gaia advances a robust, scalable data-extraction pipeline that keeps pace with rapidly evolving AI developments. Gaia expands to support additional use cases, enhancing versatility in an evolving landscape. Its architecture builds on widely available technologies to remain reliable, adaptable, and easily deployable across central banks.

 

Note to editors

BIS Innovation Hub projects are experimental in nature, for the purpose of investigating technological and practical feasibility.