In practice, AI doesn’t usually show up as a headline like “robot caused harm.” Instead, it often appears indirectly in the workflow and chart.
In the Godfrey area, we commonly see concerns such as:
- Inconsistent operative or progress notes (details that don’t match the timeline of care)
- Automated summaries or templated documentation that omits key safety steps
- Imaging interpretation that appears to rely on software outputs rather than clearly documented clinical verification
- Decision-support references in the chart—risk scores, alerts, or generated recommendations—that raise questions about supervision
If you’ve noticed any of these patterns, don’t assume it’s “just how records are written.” In these cases, the timing and context matter—especially when the outcome is severe.


