EU AI Act - May 2026: High-risk deadline extended to Dec 2, 2027 (Digital Omnibus); obligations unchanged.

Read the update →

Industry — Fintech

Credit scoring and fraud detection AI is explicitly listed in EU AI Act Annex III. Your US-incorporated fintech is not exempt.

If your product makes credit decisions, assesses financial risk, or generates lending recommendations for EU consumers, the EU AI Act's high-risk provisions apply to your system.

Fundamental Rights Impact Assessment

High-risk AI systems in financial services must conduct a fundamental rights impact assessment before deployment. This evaluates systemic bias, discriminatory outcomes, and impacts on protected characteristics.

EU AI Act — Article 27

Accuracy and Robustness Requirements

High-risk AI systems must achieve appropriate levels of accuracy and be resilient to errors. You must document performance benchmarks and establish mechanisms for identifying and correcting errors.

EU AI Act — Article 15

Data Governance Obligations

Training data for high-risk AI systems must meet specific quality criteria. Bias testing, data provenance documentation, and ongoing data governance processes are required.

EU AI Act — Article 10

Common Findings

What Belto identifies in fintech AI compliance scans

01

Fundamental rights impact assessment not conducted

High-risk AI systems in financial services deployed by public bodies or services of general interest must conduct a fundamental rights impact assessment before deployment. Most fintech teams are unaware this obligation exists.

EU AI Act Article 27

02

Training data governance documentation absent

AI systems used in credit scoring must meet specific data governance requirements. Training data must be relevant, representative, and free from errors. Documentation of data sources, selection criteria, and bias testing is required.

EU AI Act Article 10

03

No accuracy benchmarks documented

High-risk AI systems must achieve appropriate levels of accuracy, robustness, and cybersecurity. Providers must specify accuracy metrics and demonstrate performance against them. Most fintech AI deployments have not documented these benchmarks in compliance-ready form.

EU AI Act Article 15

04

Consumer transparency notice missing

Individuals subject to AI-assisted credit decisions have the right to be informed. The system's role, the logic applied, and their right to a human review must be documented and accessible.

EU AI Act Article 13 · GDPR Article 22

Deadlines

Compliance timeline for fintech AI operators

February 2025

AI Literacy obligations active

All providers and deployers must ensure staff involved in operating AI systems have adequate AI literacy. This includes credit analysts, underwriters, and risk teams using AI-assisted tools.

December 2027

Annex III high-risk financial AI obligations

Following the Digital Omnibus timeline, the deadline for full application of EU AI Act requirements for Annex III high-risk systems — including credit scoring, creditworthiness assessment, and financial risk AI — is December 2, 2027. Conformity assessments, technical documentation, fundamental rights impact assessments, and human oversight must be in place by then. Article 5 prohibited practices and Article 4 AI literacy obligations have applied since February 2025.

Ongoing

GDPR automated decision rights active

Individuals have the right not to be subject to solely automated decisions producing legal effects. Financial services companies must maintain human review mechanisms and respond to data subject requests within statutory timeframes.

Belto identifies every compliance obligation for your credit and fraud AI systems across EU, UK, and Canadian financial regulation.

From candidate ranking to performance assessment, Belto identifies exactly what your product needs before your EU customers ask.