AI generated notes as evidence

AI-powered note-takers are quickly becoming an important productivity tool.

While these tools offer efficiency and convenience, risks may arise when AI-generated notes are produced in regulatory investigations or commercial disputes.

Can AI-generated notes be used as evidence?

In regulatory investigations and legal proceedings, AI generated notes may be required to be produced and subsequently used as evidence of what occurred during a meeting.

This includes situations where informal meetings later become central to disputes, investigations or enforcement action.

AI generated records may be particularly probative where the output appears be a verbatim record of a meeting, containing significantly more detail than informal meeting notes.

Where AI note-takers can expose organisations to risk

In investigations and disputes, details matter. While an AI note-taker may appear to create a complete or verbatim meeting record, such a record could be inaccurate or incomplete, for instance, because the AI note-taker:

• misattributes statements to particular speakers
• inaccurately captures complex or nuanced discussions
• fails to record non-verbal communication, tone or context
• produces summaries that appear authoritative but are incomplete or inaccurate

Should this meeting be recorded by AI at all?

When deploying an AI note-taker, organisations should actively consider whether the tool is appropriate for the meeting in question.

Relevant considerations include whether the meeting involves:

• sensitive or confidential matters
• legally privileged information
• discussions with competitors
• potential disputes or investigations
• significant commercial or strategic decisions

Even where discussions appear routine, caution is warranted. Regulatory investigations and disputes frequently arise from interactions that seemed innocuous at the time. An AI generated record may be difficult to contextualise or explain after the fact, particularly where a significant period of time has lapsed between the generation of the record and the commencement of the investigation or dispute.

When using AI note-takers, what does 'good' look like?

Where AI tools are used to prepare transcripts or minutes, governance and accuracy controls are essential, particularly for sensitive meetings.

General principles of good note-taking continue to apply:

1. Accuracy and appropriate detail: Records should faithfully reflect what occurred at an appropriate level of detail.
2. Review and confirmation: Where possible, meeting participants should review AI-generated notes and confirm the accuracy of the record and correct errors or omissions.
3. Finalisation after review: AI-generated notes should be clearly marked as draft until reviewed and approved.

These steps help reduce the risk that inaccurate or incomplete AI-generated records are later relied upon in ways that were never intended.

What this means for businesses

• AI note-takers create records that may later be relied on in investigations or disputes
• default use of AI note-takers without governance increases organisational risk
• clear guidance on when, how and whether AI note-takers are used, is now part of good governance

Part of Lander & Rogers’ AI in Practice series - practical legal insight on AI risk, regulation and readiness for boards and decision-makers.