The Columbia study's most consequential finding is not that AI scribes make mistakes — it is that liability for those mistakes falls on nurses while the tools themselves remain unvalidated and unpoliced. Nurses are legally responsible for what appears in the chart, yet the AI-generated text they are accountable for was never subject to the regulatory scrutiny applied to other clinical decision-support software. The AI charting tools study's racial bias findings expose a structural mismatch: deployment has outrun validation, and the institution absorbing that risk is the bedside clinician, not…