Your AI Notetaker Has Receipts
A printable reference with the 16 decisions and the guidance behind each — plus three things the legal guidance leaves out.

AI notetakers crept into corporate existence. The tools are cheap, they work, and switching one on does not require anyone's approval, so adoption ran ahead of policy.
What the firms writing on this recommend — Mayer Brown, Lockton, Faegre Drinker, Foley & Lardner, White & Case, and ICO and FTC guidance — is now in one printable reference: 16 questions, the setting to choose for each, and the basis for it. The guide is here.
3 things worth adding:
Value is fast; liability is slow
Almost all of a transcript's usefulness is spent in the first week: the recap, the action items, the thing you half-remember from Tuesday. Its exposure accrues for as long as it exists, and grows as the context around it fades.
That asymmetry is the argument for short retention, far stronger than a compliance argument. A 90-day compromise does not buy 90 days of value. It buys 1 week of value and 83 days of inventory.
Inconsistency is the evidentiary problem
The instinct is to leave it to judgment: record when it seems useful, skip when it seems sensitive. That produces the worst version. When some meetings have transcripts and others do not, the gaps get read as intent, and the person deciding was the meeting organizer with no view of why it may matter later.
A rule that says which categories of meeting are captured is defensible. 15 people exercising discretion is a pattern that gets interpreted out of context.
The unpriced cost is candor
Boards already know this, which is why minutes are curated. The same dynamic applies to a discovery call, a vendor negotiation and a post-incident review: people who know they are producing a written record produce a different conversation.
The meeting where the truth goes unsaid costs more than the time the recap saved.
The settings are the policy
The guide has 16 common questions about AI recordings with the latest legal citations for each.
The default settings are the policy until someone decides otherwise.
Sources
- AI Notetakers: Productivity Tool or Emerging Legal Risk? — Mayer Brown, June 2026.
- AI notetaking and transcription: risk management considerations and best practices — Lockton, October 2025.
- Permission to Record: Considerations for AI Meeting Assistants — Faegre Drinker, February 2025.
- For Your Eyes Only? Not Quite: Shadow AI in the Workplace — Foley & Lardner, April 2026, with its June 2026 guidance on committee meetings.
- When every word is recorded: AI meeting tools and the new governance risks — White & Case, November 2025.
- Principle (e): Storage limitation — Information Commissioner's Office.
- AI Companies: Uphold Your Privacy and Confidentiality Commitments — Federal Trade Commission, January 2024.