The Camera Points Both Ways
It's 3 a.m. A med tech catches her own near miss. The wrong cup, caught at the door. The ambient system already logged the timestamp. Her pen is over the incident form. Does she write it up, or wait to see if anyone asks?
That hesitation now has a number. Black Book Research surveyed 202 nurses in August. 33% said they're less willing to report a near miss. 65% said they feel watched or tracked by at least one AI system. Only 38% feel safe overriding it (TechTarget).
It's a small sample, and these aren't senior living staff. Read it as an early signal, not a verdict on our sector. But the tools are the same ones moving into our buildings.
Home care is seeing it too. A peer-reviewed study published September 22 interviewed 43 home care stakeholders: aides, agency leaders, worker advocates, clinicians, and technology companies. They flagged surveillance and added administrative burden as core risks. The authors call for governance structures "before these technologies become fully entrenched" (Journal of General Internal Medicine).
Put that next to ECRI's finding that 35% of 124 surveyed quality, safety, and risk leaders didn't know whether they'd encountered an incorrect or misleading AI output in the past year (ECRI).
Leaders can't see the errors. Staff have stopped volunteering them.
It's not an adoption problem. It's a reporting culture problem.
Three moves this month. Publish a one-page notice telling staff what your AI systems capture about them and whether any of it feeds performance reviews. Add an AI line to your near-miss form. Then adopt the Safe Override Rule: no one is ever penalized for reporting an AI near miss, or for overriding the AI when their judgment says to.
Read the full research behind this story, plus this week's insurance coverage and governance flagships, in the LITT Briefing on Substack.