A.I. note takers are making lawyers nervous

· Source ↗

TLDR

  • AI meeting note-takers are raising attorney-client privilege concerns and creating unprecedented volumes of discoverable corporate records.

Key Takeaways

  • AI bots routinely join calls and upload full transcripts to third-party SaaS infrastructure, outside standard recording-consent workflows.
  • Attorney-client privilege may be voided when an AI note-taker is present during otherwise protected legal conversations.
  • Casual internal conversations that previously left no record are now permanently documented and subject to legal discovery.
  • Default deployment patterns give meeting participants no clear signal that transcription and summarization are active.

Hacker News Comment Review

  • Commenters flagged that AI summaries are not reliable transcripts: Gemini-style tools routinely misattribute facts, creating a defense surface but also a liability when the error goes unnoticed.
  • A split emerged between those who want local or consent-gated transcription and those resigned to assuming all digital communication is recorded, echoing one-party consent state norms.
  • The accuracy gap cuts both ways legally: note-takers cannot be cross-examined, their prompting is opaque, and small hallucinations can contradict witness testimony.

Notable Comments

  • @samuelknight: AI notes are inadmissible-adjacent – opaque prompting and inability to cross-examine create real evidentiary challenges for both sides.
  • @coffeebeqn: “Gemini said we’re rolling out our payment setup in Russia” – a live example of a hallucination with serious compliance implications.

Original | Discuss on HN