The EBA is the EU body translating the AI Act into banking practice. It published a factsheet on the Act's implications for banking and payments in November 2025, reported on AI adoption across EU banks the same year, and through 2026–27 is coordinating a common supervisory approach among national authorities while feeding the banking perspective into the EU AI Office. Its earlier machine-learning-for-IRB work already governs how ML enters regulatory capital models.
| Full name | European Banking Authority |
| Role | EU banking standard-setter |
| Force on banks | Supervisory guidance |
| Applies to | All EU credit institutions and payment providers, via national competent authorities |
| Key document | Factsheet on the AI Act's implications for banking and payments (Nov 2025) |
| Latest move | 2026–27 workplan: supervisory convergence on AI Act implementation in banking |
The EBA has worked on bank AI longer than most: a 2020 report on big data and advanced analytics set early governance expectations, and its 2021–23 workstream on machine learning for internal ratings-based (IRB) models defined how banks may use ML in capital models without losing explainability and supervisability.
With the AI Act now applicable to high-risk banking use cases, the EBA's role has shifted to implementation: mapping AI Act duties onto the existing CRD/CRR and guidelines framework so banks are not regulated twice or inconsistently, promoting supervisory convergence among national competent authorities, and participating in the AI Board's financial-services subgroup. Its 2025 adoption report — covering credit scoring, fraud detection, customer support, and general-purpose AI — is the best official snapshot of what EU banks actually deploy.
| Date | Type | Document / event |
|---|---|---|
| Jan 1, 2020 | Report | Report on big data and advanced analytics. Early EU expectations for bank use of data-driven models: governance, explainability, consumer protection. |
| Nov 1, 2021 | Consultation | Discussion paper on machine learning for IRB models. Opens the question of how ML can be used in regulatory capital models; followed by 2023 follow-up setting expectations. |
| Sep 1, 2025 | Report | Report: rising application of AI in EU banking and payments. Official snapshot of AI across EU banks — 92% already deploying AI, led by credit scoring, fraud detection, and customer support, with general-purpose and agentic AI under monitoring. |
| Nov 21, 2025 | Guidance | Factsheet: AI Act implications for banking and payments. Maps AI Act obligations (especially high-risk credit scoring) onto the existing EU banking framework; finds the Act complementary, with no significant contradictions. |
| Jan 1, 2026 | Framework | 2026–27 AI Act implementation workplan. Supervisory-convergence activities among national competent authorities, input to the European AI Office, and participation in the AI Board Subgroup on Financial Services. |
- EBA guidance reconciling AI Act model requirements with existing EBA guidelines (loan origination, internal governance)
- Supervisory-convergence output for national authorities enforcing the high-risk regime on banks
- Whether ML-for-IRB expectations get updated for the AI Act era
What has the EBA said about the AI Act for banks?
Its November 2025 factsheet explains how AI Act obligations — particularly the high-risk classification of credit scoring — interact with the existing EU banking framework, and its 2026–27 workplan focuses on getting national supervisors to apply the Act to banks consistently.
Can EU banks use machine learning in capital (IRB) models?
Yes, within the expectations set by the EBA's machine-learning-for-IRB workstream (2021 discussion paper and 2023 follow-up): banks must preserve explainability, human understanding of model drivers, and supervisability. The ECB applies these expectations in model approvals.
How common is AI in EU banking?
The EBA's 2025 adoption report found AI in production across most EU banks — dominated by fraud detection, credit scoring, and customer support, with general-purpose AI spreading fast in internal productivity use cases.
Follow every move these regulators make
6 curated AI stories for banking executives · Every morning · Free
Subscribe to BankingNewsAI →