# Central Bank of Ireland Regulatory & Supervisory Outlook 2026: Regulatory & Supervisory Outlook 2026 (Spotlight 1: Approaching AI from a Supervisory Perspective)

Source: https://www.bankingnewsai.com/ai-regulation/documents/cbi-regulatory-supervisory-outlook-2026
Last updated: Nov 19, 2026

The Central Bank of Ireland's Regulatory & Supervisory Outlook 2026, published on 26 February 2026, judged that risks associated with data, models and AI had increased since the previous year and set out four standards firms must adhere to when deploying AI: strategic alignment, accountability and explainability, proportionate governance, and compliance with all EU AI Act obligations. Its Spotlight 1, 'Approaching AI from a Supervisory Perspective', warns that the safeguards built into providers' AI models 'can be brittle' and calls agentic AI 'one frontier AI risk'. For banks, the 2026 work plan includes an assessment of the AI landscape in the sector and support for SSM work on banks' use of generative AI.

## At a glance

| Field | Value |
| --- | --- |
| Authority | [Central Bank of Ireland](https://www.bankingnewsai.com/ai-regulation/ireland) |
| Type | Report |
| Status | Final |
| Published | Feb 26, 2026 |
| Applies to | All firms regulated by the Central Bank of Ireland, including banks, credit unions, payment and e-money firms, insurers and investment firms. A risk outlook and statement of supervisory priorities: its four AI standards are supervisory expectations, while AI Act obligations are binding law in their own right |
| Official text | https://www.centralbank.ie/docs/default-source/publications/regulatory-and-supervisory-outlook-reports/regulatory-supervisory-outlook-report-2026.pdf |

## Key points

- Four standards ('Firms must adhere to the following standards when deploying AI'): strategic alignment, so that AI use 'is appropriate for the specific business challenge being addressed'; accountability and explainability, with 'clear accountabilities and responsibility, human oversight of decisions and their explainability'; proportionate governance, 'commensurate with the scale, scope and sensitivity of the AI deployment'; and compliance with all EU AI Act obligations, including transparency requirements.
- Risk assessment: data, modelling and AI risks are one of the report's risk themes; widespread adoption of third-party AI tools 'changes the risk landscape and calls for stronger model governance, data quality, transparency and accountability', with model risk 'a growing concern'.
- Agentic AI: agents are being deployed in some firms' internal operations; they add 'system complexity which increases the challenge of understanding and managing risk', and agent-based transactions need further monitoring for conduct, financial integrity and operational resilience risk.
- AI supply chains: disruptions from trade or technology restrictions 'could impair access to critical AI services', and reliance on AI and cloud providers from a small number of locations requires firms to consider contingency options.
- Consumer risk and safeguards: methods to assess AI conduct risk remain 'significantly underdeveloped'; Box 4 reports AIR-BENCH 2024 results showing wide variation in how models refuse harmful requests on insurance eligibility, scams and creditworthiness, and concludes there is 'a pressing need for robust safety evaluations by both deploying firms of their own safeguards and by supervisors implementing the AI Act'.
- Supervisory stance: the approach is 'consistent with that being taken by the EU AI Office', and the ESAs are converging on an outcome-focused approach anchored in the AI Act, DORA and sectoral rules, 'not treating AI as a silo'.
- AI Act: 'The Government has designated the Central Bank as a Market Surveillance Authority (MSA) under the AI Act for high-risk AI system use cases in the financial services sector'; the Central Bank supports the Commission's November 2025 simplification proposal, which would move the start of high-risk requirements to December 2027.
- Banking work plan for 2026: 'Assessment of the AI landscape in the banking sector to build supervisory knowledge on AI use', and support for SSM work 'to strengthen supervisory understanding of how banks use generative AI applications'.

## What changed for banks

The 2026 Outlook is the first to state AI standards that firms 'must adhere to', and it raised the Central Bank's assessment of data, model and AI risk. It adds agentic AI, AI supply-chain resilience and the brittleness of model safeguards to the 2025 risk picture, and it reports the Central Bank's AI Act designation as done rather than expected. It remains a supervisory statement: the standards are expectations, while the AI Act's own obligations are binding.

## Use cases it governs

- [AI governance (general)](https://www.bankingnewsai.com/ai-regulation/by-use-case#governance-general)
- [Model risk management](https://www.bankingnewsai.com/ai-regulation/by-use-case#model-risk)
- [Generative & agentic AI](https://www.bankingnewsai.com/ai-regulation/by-use-case#generative-agentic-ai)
- [Third-party & vendor AI](https://www.bankingnewsai.com/ai-regulation/by-use-case#third-party-vendors)
- [Cybersecurity](https://www.bankingnewsai.com/ai-regulation/by-use-case#cybersecurity)
- [Credit scoring & underwriting](https://www.bankingnewsai.com/ai-regulation/by-use-case#credit-underwriting)

## FAQ

### What are the Central Bank of Ireland's AI expectations for firms?

The Regulatory & Supervisory Outlook 2026 lists four standards firms must adhere to when deploying AI: strategic alignment (AI must suit the specific business challenge), accountability and explainability (clear responsibility, human oversight of decisions and their explainability), proportionate governance (risk management commensurate with the scale, scope and sensitivity of the deployment) and compliance (processes to meet all EU AI Act obligations, including transparency).

### What does the Central Bank of Ireland's 2026 Outlook say about agentic AI?

It says agents are being deployed in some financial firms' internal operations, that it is not clear exactly how agentic AI will develop, and that it adds system complexity that makes risk harder to understand and manage. It calls agentic AI one frontier AI risk and says agent-based transactions need further monitoring for conduct, financial integrity and operational resilience risk.

## Related documents

- [Governor Makhlouf speech on AI and tokenisation (2 November 2026)](https://www.bankingnewsai.com/ai-regulation/documents/cbi-makhlouf-financial-system-conference-2026) — Economic Transitions: Technology (opening remarks by Governor Makhlouf at the Financial System Conference) (Nov 2, 2026)
- [Regulation of Artificial Intelligence Act 2026 (Ireland)](https://www.bankingnewsai.com/ai-regulation/documents/ireland-regulation-of-ai-act-2026) — Regulation of Artificial Intelligence Act 2026 (No. 31 of 2026) (Jul 21, 2026)
- [S.I. No. 366/2025 AI Act designation regulations](https://www.bankingnewsai.com/ai-regulation/documents/ireland-si-366-2025-ai-designation) — European Union (Artificial Intelligence) (Designation) Regulations 2025 (Jul 25, 2025)
- [Central Bank of Ireland Regulatory & Supervisory Outlook 2025](https://www.bankingnewsai.com/ai-regulation/documents/cbi-regulatory-supervisory-outlook-2025) — Regulatory & Supervisory Outlook 2025 (Spotlight 2: AI: Balancing Opportunities and Risks) (Feb 28, 2025)
- [Colm Kincaid speech on AI in finance (22 May 2024)](https://www.bankingnewsai.com/ai-regulation/documents/cbi-kincaid-ai-consumer-protection-speech-2024) — Financial consumer protection and market conduct considerations of AI in finance (May 22, 2024)
- [Central Bank of Ireland Regulatory & Supervisory Outlook 2024](https://www.bankingnewsai.com/ai-regulation/documents/cbi-regulatory-supervisory-outlook-2024) — Regulatory & Supervisory Outlook 2024 (Spotlight 2: Artificial Intelligence: A Supervisory Perspective) (Feb 29, 2024)
- [BNM AI Feedback Statement](https://www.bankingnewsai.com/ai-regulation/documents/bnm-ai-feedback-statement-2026) — Feedback Statement: Artificial Intelligence (AI) in the Malaysian Financial Sector (Nov 19, 2026)
- [BoE/PRA/FCA effective practices on AI cyber risk (Nov 2026)](https://www.bankingnewsai.com/ai-regulation/documents/uk-effective-practices-cyber-risk-ai-2026) — Effective practices: Cyber risk and resilience management relating to artificial intelligence (AI) (Nov 19, 2026)

Last reviewed Nov 19, 2026. Cite the official text (https://www.centralbank.ie/docs/default-source/publications/regulatory-and-supervisory-outlook-reports/regulatory-supervisory-outlook-report-2026.pdf) for the rule and this page for the summary and dates.

---

Canonical page: https://www.bankingnewsai.com/ai-regulation/documents/cbi-regulatory-supervisory-outlook-2026
Part of [BankingNewsAI](https://www.bankingnewsai.com/) — a free daily brief on AI in banking, an AI regulation tracker (60 authorities, 360 documents) and AI-strategy profiles of the 120 largest US banks. Markdown versions of every reference page: append `.md` to the page URL; index at https://www.bankingnewsai.com/llms.txt.
