Bewaard in:
Bibliografische gegevens
Hoofdauteurs: Clarke, Laura, Nemade, Kshitij, Adamson, Jack, Joannides, Alexis
Formaat: Recurso digital
Taal:Engels
Gepubliceerd in: Zenodo 2026
Onderwerpen:
Online toegang:https://doi.org/10.5281/zenodo.19255946
Tags: Voeg label toe
Geen labels, Wees de eerste die dit record labelt!
_version_ 1866901797883346944
author Clarke, Laura
Nemade, Kshitij
Adamson, Jack
Joannides, Alexis
author_facet Clarke, Laura
Nemade, Kshitij
Adamson, Jack
Joannides, Alexis
contents <div> <p><span lang="EN-GB">This report outlines how the Eastern England Secure Data Environment (EE</span><span lang="EN-GB">‑</span><span lang="EN-GB">SDE) has been preparing to support safe and trustworthy use of artificial intelligence (AI) and machine</span><span lang="EN-GB">‑</span><span lang="EN-GB">learning (ML) models trained on sensitive health data. As part of VISTA, one of DARE UK’s Early Adopter projects, the EE</span><span lang="EN-GB">‑</span><span lang="EN-GB">SDE deployed and evaluated new tools designed to help Trusted Research Environments (TREs) assess and manage privacy risks linked to AI projects. </span> <br><br></p> </div> <div> <p><span lang="EN-GB">A key outcome of this work is the VISTA AI Risk Assessment Toolkit, which provides a structured way for TREs to review proposed AI projects, understand their data needs, and ensure that appropriate safeguards are in place from the outset. The toolkit works alongside SACRO</span><span lang="EN-GB">‑</span><span lang="EN-GB">ML, a new disclosure</span><span lang="EN-GB">‑</span><span lang="EN-GB">control technology that checks trained ML models for signs that they might reveal information about individuals in the training data. Together, these tools help reviewers make clearer, evidence</span><span lang="EN-GB">‑</span><span lang="EN-GB">based decisions about whether a model can safely be released. </span> <br><br></p> </div> <div> <p><span lang="EN-GB">The project shows that AI safety checks can be integrated into real research workflows without disrupting analysis. It also highlights areas for future improvement, including clearer guidance, better onboarding materials, and continued collaboration across the TRE community to build consistent and trusted approaches to responsible AI research.</span> </p> </div>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_19255946
institution Zenodo
language eng
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle Eastern England SDE Safe Models, Safe AI: Governance and Disclosure Testing
Clarke, Laura
Nemade, Kshitij
Adamson, Jack
Joannides, Alexis
VISTA
TREvolution
SACRO-ML
Model Disclosure Control
<div> <p><span lang="EN-GB">This report outlines how the Eastern England Secure Data Environment (EE</span><span lang="EN-GB">‑</span><span lang="EN-GB">SDE) has been preparing to support safe and trustworthy use of artificial intelligence (AI) and machine</span><span lang="EN-GB">‑</span><span lang="EN-GB">learning (ML) models trained on sensitive health data. As part of VISTA, one of DARE UK’s Early Adopter projects, the EE</span><span lang="EN-GB">‑</span><span lang="EN-GB">SDE deployed and evaluated new tools designed to help Trusted Research Environments (TREs) assess and manage privacy risks linked to AI projects. </span> <br><br></p> </div> <div> <p><span lang="EN-GB">A key outcome of this work is the VISTA AI Risk Assessment Toolkit, which provides a structured way for TREs to review proposed AI projects, understand their data needs, and ensure that appropriate safeguards are in place from the outset. The toolkit works alongside SACRO</span><span lang="EN-GB">‑</span><span lang="EN-GB">ML, a new disclosure</span><span lang="EN-GB">‑</span><span lang="EN-GB">control technology that checks trained ML models for signs that they might reveal information about individuals in the training data. Together, these tools help reviewers make clearer, evidence</span><span lang="EN-GB">‑</span><span lang="EN-GB">based decisions about whether a model can safely be released. </span> <br><br></p> </div> <div> <p><span lang="EN-GB">The project shows that AI safety checks can be integrated into real research workflows without disrupting analysis. It also highlights areas for future improvement, including clearer guidance, better onboarding materials, and continued collaboration across the TRE community to build consistent and trusted approaches to responsible AI research.</span> </p> </div>
title Eastern England SDE Safe Models, Safe AI: Governance and Disclosure Testing
topic VISTA
TREvolution
SACRO-ML
Model Disclosure Control
url https://doi.org/10.5281/zenodo.19255946