Assess privilege risk before ingesting advice files into AI
MaintainedLast reviewed
Key facts
- Human involvement
- YellowClose supervision
- How AI is used
- Classify privileged advice and check the tool's data terms.
- What the lawyer must review
- Risk partner decides what may be ingested.
- Written for / version
- Written for Australia
The work product
Forum. Australia
Produce. Assess privilege risk before ingesting advice files into AI
Have ready. Clayton Utz (Apr 2026): AI workflows and privilege; Hall & Wilcox: AI and waiver of privilege; The matter instructions and the documents the output will rely on..
Example of the limit. Stop before the result is filed, sent or relied on. Risk partner decides what may be ingested.
Stop and hand it back when
- The cited source is not the instrument that applies to this forum or this date.
- The output states a fact, date, citation, figure or commitment that is not in the instructions or the source.
- The responsible lawyer has not approved the exact output for its intended use.
Sources for this task were checked on 2026-10-10.
Who does what
| Role | What they do |
|---|---|
| Assistance | Classify privileged advice and check the tool's data terms. |
| Lawyer | Set the scope, supervise the work closely, verify every part relied on, resolve omissions and exceptions, and approve the exact output. |
Checklist
- Open the cited source and confirm it is the current instrument for this forum and this date.
- Risk partner decides what may be ingested.
- Compare the output with the source and the instructions, including material the output left out.
- The responsible lawyer approves the exact output for its intended use, or sends it back.
Sources
- Clayton Utz (Apr 2026): AI workflows and privilege · 2026-10-10
- Hall & Wilcox: AI and waiver of privilege · 2026-10-10
Suggested working checks from the linked public sources. Read the source and confirm it applies to this forum and this date. The checks do not approve software or record that the work is done.