Precedents: Confidentiality & client information

5 starting-point precedents for this area. Each one names the law it rests on and says how far AI may help.

Precedents in this area
5
Full text
For signed-in accounts
Newest law stated as at
17 October 2026
Guidance for this area
Tasks and AI use in Confidentiality & client information
Precedents for Confidentiality & client information
PrecedentKindWhere it appliesAI useLaw stated as at
AI confidentiality rules for staff (short form of the firm AI policy)Short-form confidentiality rules drawn from the firm's master AI use policy: which information classes may go into which tools, court-restricted material, privilege and barriers, client choices, note-takers and agents, and incident reporting, with an upload checklist.PolicyAustralia, New South Wales, Victoria Yellow17 October 2026
Staff and contractor AI confidentiality undertaking and acknowledgementA signed undertaking by each lawyer, staff member and contractor to follow the firm's AI confidentiality controls: approved tools only, no personal accounts, no bypassing controls or barriers, verification, incident reporting, monitoring and duties after leaving.FormAustralia Yellow17 October 2026
AI vendor confidentiality and privilege questions (short form of the vendor due-diligence questionnaire)Short-form vendor questions on confidentiality only: training and other use of firm data, retention and deletion, location and sub-processors, matter access and information barriers, court conditions for restricted material, confidentiality terms and incident notice.ChecklistAustralia, New South Wales Yellow17 October 2026
Engagement clause: client authority for AI processing of confidential information (rider to the AI engagement clauses)Confidentiality rider for the engagement letter: the firm's safeguards for AI processing, the client's express, limited or refused authority under ASCR r 9.2.1, and overseas processing. Use with the AI engagement and costs clauses, not instead of them.ClauseAustralia Yellow17 October 2026
AI data incident steps (confidentiality short form of the AI incident response procedure)Confidentiality steps for AI incidents that expose information (public chatbots, wrong workspaces, vendor leaks, misdirecting agents, restricted court material, note-taker bots), with the privilege, privacy, court-order and barrier assessment. Runs inside the master procedure.PolicyAustralia, New South Wales Yellow17 October 2026

AI use levels: Green means AI can do most of the drafting and a lawyer reviews it. Yellow means AI helps in set ways and a lawyer checks specific points. Red means a lawyer does the work and AI stays out of the decisive steps.

All precedents