Lorenzo’s Portfolio
Store
← Articles & research

Research paper

Keep It Secret, Keep It (in a) Safe, Then Cast It into the Fire: Engineering AI Act Data Governance — A Practical Guide for Data Teams

Lorenzo Colombani

Maps the AI Act’s data obligations, article by article, onto Data Vault design decisions: sensitive satellites, where the bias gate sits, which erasure pattern fits, what evidence may survive a deletion. 24 pages.

Read paper
About this paperReading notes

About the research

Preserve the history. Know when to erase it.

This preprint examines the tension between Data Vault’s preservation of history and data-protection duties that can require deletion. A recruitment-screening example makes the architectural questions concrete.

It maps selected AI Act data obligations to collection, storage, transformation, release, use and deletion decisions. Proposed patterns connect requirements with evidence, an accountable owner and a test.

By
Lorenzo Colombani
Format
Research paper
Published
2 September 2026
DOI
10.5281/zenodo.22255574
Licence
CC BY 4.0
Preview
content visualization

Reading notes

Preserve the history. Know when to erase it.

The architectural tension

Sensitive attributes can matter for bias testing while also requiring strict controls. The paper explores where those attributes sit, where a bias gate operates and which erasure pattern fits.

The scope of the proposal

The paper presents legal and architectural analysis with suggested engineering practices. It distinguishes those patterns from a legal basis and from the Act’s broader organisational requirements.

Preview this work

More articles & research →