On 14 October 2025 Pedro Machado, ECB representative to the Supervisory Board, spoke in Lisbon at Banco de Portugal on 'Artificial intelligence and supervision: innovation with caution'. He named five ECB supervisory AI tools — Agora, Athena, Delphi, Medusa and Heimdall — and stated that AI 'does not replace human judgement'. He flagged five risks: hallucinations ('wrong but confident'), deskilling of staff, lack of explainability, cyber and operational vulnerabilities, and 'AI talking to AI' as banks draft submissions with AI that supervisors then read with AI.
| Document | Machado speech: 'Artificial intelligence and supervision: innovation with caution' (Oct 2025) — Artificial intelligence and supervision: innovation with caution — speech by Pedro Machado |
| Issued by | European Central Bank — Banking Supervision (SSM) |
| Type | Speech |
| Status | Final |
| Published | Oct 14, 2025 |
| Applies to | No obligations; explains how the ECB uses AI in supervision and what it worries about when banks use AI in their supervisory reporting. |
| Official source | bankingsupervision.europa.eu ↗ |
| Use cases | Generative & agentic AI · AI governance (general) |
What are the key points of Machado speech: 'Artificial intelligence and supervision: innovation with caution' (Oct 2025)?
- Five named ECB tools: Agora (data lake), Athena (document search and summarisation with visible sources), Delphi (emerging-risk detection from market and social-media signals), Medusa (findings and measures repository), Heimdall (fit-and-proper support)
- ECB adopts AI 'carefully, transparently and with clear use cases'; human judgement remains decisive
- Risk of opacity when banks use AI to draft what they send supervisors and supervisors use AI to review it
- Five risks: inaccuracy/hallucination, deskilling, explainability gaps, cyber and operational vulnerabilities, AI-to-AI loops
- Frames ethics and conduct of supervisors as the safeguard around AI-assisted supervision
What did Machado speech: 'Artificial intelligence and supervision: innovation with caution' (Oct 2025) change for banks?
It was the clearest public account of which AI systems the ECB runs and how it thinks about their limits, and it put banks on notice that AI-generated regulatory submissions will be scrutinised for substance behind the polish.
Does the ECB use generative AI to review bank submissions?
Yes — Athena and related tools search, translate and summarise supervisory documents — but Machado's October 2025 speech stresses the tools inform rather than replace supervisory judgement.
| Date | Document | Status |
|---|---|---|
| Jul 7, 2026 | ECB 'Dear CEO' letter on AI-enabled cybersecurity threats (SSM-2026-0301) — Addressing AI-enabled cybersecurity threats — letter from the Chair of the Supervisory Board to CEOs of significant institutions | In force |
| Jun 3, 2026 | Elderson speech: 'Strengthening operational resilience for the age of AI' (June 2026) — Strengthening operational resilience for the age of AI — speech by Frank Elderson | Final |
| Feb 24, 2026 | Machado speech: 'Technology is neutral, governance is not' (Feb 2026) — Technology is neutral, governance is not: AI adoption in the banking sector — speech by Pedro Machado | Final |
| Feb 3, 2026 | Montagner speech: 'Encouraging innovation, managing risks' (Feb 2026) — Encouraging innovation, managing risks: the ECB's approach to digital transformation — speech by Patrick Montagner | Final |
| Nov 20, 2025 | Supervision Newsletter: AI use cases for credit scoring and fraud detection (Nov 2025) — AI's impact on banking: use cases for credit scoring and fraud detection (Supervision Newsletter, November 2025) | Final |
| Nov 18, 2025 | SSM supervisory priorities 2026–28 — ECB Banking Supervision: SSM supervisory priorities for 2026–28 | In force |
Follow every document these regulators publish
6 curated AI stories for banking executives · Every morning · Free
Subscribe to BankingNewsAI →