High-risk requirements: August 2026

Regulation

EU AI Act — AI risk regulation for financial services

The EU AI Act is a risk-based framework that imposes binding obligations on developers and deployers of AI systems. Financial institutions that use AI in credit decisions, fraud detection or customer interactions are directly in scope.

What is the AI Act?

Regulation (EU) 2024/1689 — the EU Artificial Intelligence Act — is the world's first comprehensive legal framework for AI. It classifies AI systems by risk level and imposes proportionate obligations on providers (developers) and deployers (users of AI systems built by others).

Financial institutions are primarily deployers. If your organisation uses AI systems built by vendors for high-risk use cases, the compliance obligations fall on you as the deployer.

Risk classification and financial services

Use caseRisk levelObligations
Credit scoring / creditworthiness assessmentHigh riskFull high-risk requirements
Insurance premium calculationHigh riskFull high-risk requirements
Fraud detectionLimited/MinimalTransparency obligations only
Customer chatbotsLimitedDisclosure required
AML transaction monitoringUnder assessmentGuidance expected
Prohibited: real-time biometric surveillanceProhibitedNot permitted

Key obligations for high-risk AI deployers

Due diligence on AI systems

Before deploying a high-risk AI system, you must verify the provider has met their obligations: technical documentation, CE marking, EU declaration of conformity and registration in the EU database.

Human oversight

Deployers must implement appropriate human oversight measures. For credit decisions, this means human review procedures, override capabilities and clear accountability structures.

Data governance

Training data used in high-risk AI must meet quality standards: relevant, representative, free of errors and sufficiently complete. Deployers must ensure data governance practices meet this bar.

Monitoring and incident reporting

High-risk AI systems must be monitored for performance post-deployment. Serious incidents must be reported to national supervisors.

Timeline

DateMilestone
August 2024AI Act entered into force
February 2025Prohibited AI practices banned
August 2025GPAI model obligations apply
August 2026High-risk AI system requirements apply — financial institutions in scope
August 2027High-risk AI in regulated products (Annex I) apply

How Arcens helps

Horizon scanning: The AI Act's technical standards and guidance notes are still being developed. We track EBA, EIOPA and European Commission publications and translate them into actionable requirements.

NFR management: We help you build an AI governance framework — inventory, risk classification, oversight procedures and documentation — that meets AI Act requirements and aligns with your existing risk management structure.

Deadline alert

August 2026: High-risk AI compliance required. Credit scoring, insurance pricing AI in scope.

Preparation should begin now — AI inventory and gap assessment typically take 3–6 months.

Quick facts

Full name: Regulation (EU) 2024/1689

Supervisor (NL): TBD — likely DNB/AFM

Key deadline: August 2026

Related regulations

→ DORA → DNB SIRA

Is your AI use in scope?

Ask our advisor to assess whether your AI systems fall under the high-risk category and what steps you need to take before August 2026.

Assess AI Act scope