Generative AI for Bank Carbon Data Gaps

Banca d'Italia Publications

Published March 12th, 2026

Detected March 13th, 2026

Summary

Banca d'Italia published a paper exploring the use of Generative AI (GenAI) to address gaps and inconsistencies in carbon emission data reported by major Euro area banks. The research assesses GenAI's ability to supplement traditional data sources, noting its potential while also highlighting quality and transparency concerns.

View original document View source feed page

What changed

This publication from Banca d'Italia, titled "Generative AI can help fill banks' carbon data gaps," examines the effectiveness of Generative AI (GenAI) tools in addressing inconsistencies and deficiencies in the carbon emission data of major Euro area banks. The paper evaluates three GenAI tools, finding they can correlate with traditional data sources and help identify anomalies. However, the study also points out that GenAI-generated data has its own quality and consistency issues, alongside concerns regarding replicability and transparency.

The implications for financial institutions and compliance officers involve understanding the potential and limitations of GenAI in ESG reporting. While GenAI may offer a complementary data source, it is not a complete replacement for traditional methods. The paper suggests that specialized language models and improved emission reporting standards are necessary to overcome current GenAI limitations. Regulated entities should be aware of these developments as the industry explores new technologies for data analysis and reporting in environmental, social, and governance (ESG) compliance.

What to do next

Review the Banca d'Italia paper on GenAI for carbon data analysis.
Assess current internal processes for carbon emission data collection and reporting.
Monitor developments in GenAI applications for ESG compliance and data quality standards.

Source document (simplified)

N. 1003 - L'intelligenza artificiale generativa può aiutare a colmare le lacune nei dati carbonici delle banche?

Questioni di Economia e Finanza (Occasional Papers) di Cristina Angelico ed Enrico Bernardini Marzo 2026

I dati sulle emissioni carboniche indirette delle principali banche dell'area euro, che includono quelle delle imprese in cui le banche investono o cui fanno credito, evidenziano lacune e incoerenze nelle stime che variano in base alle fonti analizzate. Questo lavoro valuta la capacità di tre strumenti di intelligenza artificiale generativa (GenAI) di colmare tali carenze.

I dati ottenuti con gli strumenti di GenAI sono correlati con quelli ricavati dalle fonti tradizionali e aiutano a individuarne le anomalie. Tuttavia, presentano anch'essi problemi in termini di qualità e coerenza, e inoltre sollevano questioni di replicabilità e trasparenza. Lo sviluppo di modelli linguistici specializzati e migliori standard di reporting delle emissioni potrebbero contribuire a superare le limitazioni attuali della GenAi e a renderla una fonte di dati complementare a quelle tradizionali.