Swiss auditors find that FINMA's AI tools lighten supervisory work

The Swiss Federal Audit Office examined two FINMA systems, for press coverage and for risk estimates of small banks. It found that such AI methods support and relieve specific supervisory tasks.

  • Not rated an organization

    The audit office found that the systems help with data and analysis preparation but that FINMA does not measure their effect.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Reported benefit.
Evidence

foundstudiedreportedclaimed

An independent audit report supports the benefit.

The system’s part
Not isolated. The audit office examined the two systems' methods and use but did not compare supervisory work with and without them. Both use machine-learning methods, TF-IDF with k-means clustering and elastic-net regression.
Limits
Not yet checked.
Costs and harms
Not yet checked.
Later
Not yet checked.

Download data Cite Suggest a correction

Benefit type
Improved Knowledge Discovery
System
FINMA Pressespiegel Report (press-review clustering) and Vorhersage von Risikokennzahlen bei Kleinbanken (risk-weighted asset estimates for small banks)
Made by
FINMA Data Innovation Lab (DIL), in-house
Used by
Eidgenössische Finanzmarktaufsicht (FINMA, Swiss Financial Market Supervisory Authority)
Where
Switzerland
Stage
In routine use
Date
, when the source was published
Kind
Program
Broad benefit
Processing Information, Estimating the Quantifiable Characteristics of Products, Events, or Information
Collections
Scrutiny and findings
Last revised
31 October 2026

Sources

  1. Audit report, by authors with no tie to the system.

    Eidgenössische Finanzkontrolle (EFK). Prüfung des Einsatzes Künstlicher Intelligenz in der Aufsichtstätigkeit: Eidgenössische Finanzmarktaufsicht. EFK-25551, version with responses dated 14 October 2025, published 7 January 2026. Section 3.2, pp. 15-16.

    Die Analysen der beiden ausgewählten KI-Systeme zeigen, dass KI-basierte Verfahren geeignet sind, um spezifische Fachfunktionen in unterschiedlichen Anwendungsbereichen der Aufsicht punktuell zu unterstützen und entlasten.

Details

AI system

Workflow

The system for press coverage turned press texts into TF-IDF vectors, reduced them with principal component analysis and grouped them with k-means clustering. For small banks, elastic-net regression on historical supervisory data and on figures from larger institutions produced estimates of risk-weighted assets.

Follow-up

Conditions for replication

Supervisors use the press report only as a supplementary early-warning indicator. Results may be less stable or reliable when data are atypical.

Entry details

Sector

Financial supervision.

Purpose

FINMA's Data Innovation Lab built the two systems to support supervisors. One sorts daily press coverage of supervised banks by theme, and the other estimates risk-weighted assets for small banks under simplified reporting.

Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this record

Peter H. Vartanian. (2026). Swiss auditors find that FINMA's AI tools lighten supervisory work (AIBD-0715). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0715.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)

BibTeX Data Field definitions