Privilege and confidentiality in generative AI workflows

Generative AI (GenAI) systems store and process client data in three ways: in the model’s parameters through training and memorisation, in the context window during a live session, and in knowledge databases for retrieval-augmented generation (RAG). This article explains the three modes in terms accessible to practitioners. Each creates different and often counter-intuitive risks to confidentiality and legal professional privilege, and each calls for a specific governance response. Drawing on the first English and American decisions to address privilege and generative AI, on the orthodox authorities against which they must be read, and on recent computer science research, we then situate the analysis within the regulatory framework governing solicitors in England and Wales and within the ordinary principles of professional negligence. We argue that the standard of effective information governance (and with it the benchmark for negligence and misconduct) is changing. Although we write primarily for SRA-regulated practitioners, the analysis extends to any jurisdiction in which privilege or professional secrecy depends on demonstrable confidentiality.

Authors

Publication Details

Journal
OSF Preprints (OSF Preprints)
Published
2026-11-19
DOI
https://doi.org/10.31219/osf.io/5cq63
Primary Topic
Artificial Intelligence in Law
Type
preprint

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Privilege and confidentiality in generative AI workflows

Thomas Melham, Václav Janeček
1 links
OSF Preprints (OSF Preprints)
Artificial Intelligence in Law
preprint

Privilege and confidentiality in generative AI workflows

Thomas Melham, Václav Janeček
preprint en

Abstract

Generative AI (GenAI) systems store and process client data in three ways: in the model’s parameters through training and memorisation, in the context window during a live session, and in knowledge databases for retrieval-augmented generation (RAG). This article explains the three modes in terms accessible to practitioners. Each creates different and often counter-intuitive risks to confidentiality and legal professional privilege, and each calls for a specific governance response. Drawing on the first English and American decisions to address privilege and generative AI, on the orthodox authorities against which they must be read, and on recent computer science research, we then situate the analysis within the regulatory framework governing solicitors in England and Wales and within the ordinary principles of professional negligence. We argue that the standard of effective information governance (and with it the benchmark for negligence and misconduct) is changing. Although we write primarily for SRA-regulated practitioners, the analysis extends to any jurisdiction in which privilege or professional secrecy depends on demonstrable confidentiality.

OSF Preprints (OSF Preprints)Vol. 42(4)
Leverhulme Trust
Peace, justice, and strong institutions
Artificial Intelligence in Law
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