Generative AI incident register for US | AI Law Framework
Form · AI decision systems
Generative AI incident register for US law firms
Register of generative AI incidents for a US law firm, linking each to the incident file note, the Rule 1.4 and 1.6 decisions and any court correction, so partners can show the supervision Rules 5.1 and 5.3 require.
MaintainedWe keep this page up to date.Law stated as at
Kind
Form A form to complete, or a guide to completing an official one.
Version
1
Where it applies
United States
AI use
Yellow You and the tool draft together, with you directing. You check every statement you rely on.
Source
Written for this library
Status
Maintained. Checked on 7 November 2026 against the 5 authorities it cites, for United States.
When to use it
Use alongside the incident file note (a45-us-gai-confidentiality-incident-note) and the firm's generative AI policy. Each file note becomes one register line. Review the register when the policy is reviewed.
How to use this precedent
Check that it fits the matter: read “When to use it” and the places it applies to.
Log in to read the full text. It is free.
Adapt it to the matter, filling in each placeholder shown in [square brackets].
Use AI only as far as the Yellow level allows. “How AI may be used” says what AI may and must not do.
Before it goes out, a lawyer checks it against current law and each point under “The lawyer checks”.
How AI may be used
A tool may summarise entries and trends. A lawyer decides on client communication, any court correction and any state breach-notice duty, and closes each entry.
AI may
Summarise open entries and repeated causes
Prepare the review summary table
AI must not
Must not decide whether the client must be told or a court corrected
Must not decide whether a state data breach law applies
The lawyer checks
Whether Rule 1.4 requires telling the client
Whether Rule 3.3 requires correcting a statement to a tribunal
Whether a state breach-notification statute or the engagement letter requires notice