What professional rules say
Professional bodies are setting expectations. In Israel, the Israel Bar Association's National Ethics Committee published a preliminary opinion on the use of AI in legal practice in May 2024 (decision 60/24) and an updated opinion in September 2026. The updated opinion addresses, among other things, duties of loyalty and diligence, confidentiality and privilege, protection of information and verification of outputs. The guiding principle: professional responsibility for the work remains with the lawyer. Under the updated opinion, removing the client's name is not adequate protection before entering information into an open platform, and such use requires the client's prior, informed and explicit consent. A general clause in the fee agreement is not sufficient.
Updated opinion on the Israel Bar website (Hebrew)(opens an external website) · Preliminary opinion, 2024 (Hebrew)(opens an external website)
Law firms
The focus is on real use by partners, associates, trainees and staff: which tools enter the work, what data is entered and how outputs are checked before anyone relies on them.
- Client documents or professional information are entered into AI tools.
- Staff use personal accounts or tools that have not been reviewed.
- There are no clear rules for human review of outputs.
- The firm is considering enterprise licensing and wants to set conditions of use.
Accounting firms
The focus is on AI use in an environment holding clients' financial and business data, documents and operational data, with clear rules on inputs, vendors and output review.
- Teams use AI for drafting, summarizing, analysis or document work.
- There is no complete picture of the tools and accounts in use.
- Financial or business data is entered into tools without a common framework.
- A structured process is needed for tool selection, permissions and output review.
Five control layers
The shared framework for both firm types looks at five layers. The first step is mapping actual use; policy and training come after.
Use
Who uses AI, which tools, for which tasks and in which accounts.
Data
What data enters the tools, what needs restriction and which minimization rules exist.
Quality
How outputs, sources, calculations and conclusions are checked before professional reliance.
Governance
Who approves tools and uses, how exceptions are handled and how vendors and permissions are reviewed.
People
What staff know, how they receive guidance and how new uses are identified over time.
What the work can include
- Tool and use case register
- AI use policy
- Approved tools and conditions
- Data input rules
- Human review of outputs
- Training and periodic review
The goal: AI use that is visible, managed and controllable.
Start on your own
15 questions for professional firms
A checklist to read and discuss.
Read the questionsReadiness check
15 questions, a score and three areas to address.
Start the checkFor a full, supported mapping: AI Governance Quick Scan.