Minimum Viable Evidence Architecture for SME Providers of High-Risk AI Systems under the EU AI Act

Article 11(1) of the EU AI Act, as amended by Regulation (EU) 2026/1744, lets SMEs provide Annex IV technical documentation in a simplified manner. This research note argues that simplification changes how information is presented, not the provider’s responsibility to maintain the evidence behind it. It proposes a minimum viable evidence architecture in which each technical-file claim is linked to an evidence source, version, date, accountable owner and human review state, with missing, stale or conflicting evidence blocking approval. A prototype for Annex IV Section 1, using a fictional provider and simulated evidence, demonstrates the pattern. It does not demonstrate legal compliance. Produced as part of the AIGN Junior AI Governance Fellowship (2026), mentored by Patrick Upmann. This revised version has not been re-reviewed by AIGN. The author is developing Emetis, the software used for the prototype, as a commercial product.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-10-26
DOI
https://doi.org/10.5281/zenodo.23674779
Primary Topic
Digitalization, Law, and Regulation
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Minimum Viable Evidence Architecture for SME Providers of High-Risk AI Systems under the EU AI Act

Larissa Engelbrecht
Zenodo (CERN European Organization for Nuclear Research)
Digitalization, Law, and Regulation
article

Minimum Viable Evidence Architecture for SME Providers of High-Risk AI Systems under the EU AI Act

Larissa Engelbrecht
article en

Abstract

Article 11(1) of the EU AI Act, as amended by Regulation (EU) 2026/1744, lets SMEs provide Annex IV technical documentation in a simplified manner. This research note argues that simplification changes how information is presented, not the provider’s responsibility to maintain the evidence behind it. It proposes a minimum viable evidence architecture in which each technical-file claim is linked to an evidence source, version, date, accountable owner and human review state, with missing, stale or conflicting evidence blocking approval. A prototype for Annex IV Section 1, using a fictional provider and simulated evidence, demonstrates the pattern. It does not demonstrate legal compliance. Produced as part of the AIGN Junior AI Governance Fellowship (2026), mentored by Patrick Upmann. This revised version has not been re-reviewed by AIGN. The author is developing Emetis, the software used for the prototype, as a commercial product.

Zenodo (CERN European Organization for Nuclear Research)
Openalex Percentile: Top 3%
Digitalization, Law, and Regulation
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon