Case Study 4: A Clinical Safety Benchmark of Five AI Models
Five fictional patient scenarios, each with a hidden safety trap. With the rubric visible, every model looked flawless. Without it, the cracks showed.
Read more →Blog
AI field notes & case studies
How AI models behave when the stakes are real, plus updates from MixedBagPublications.
Welcome: Why a Nurse Is Testing AI
Fifteen years at the bedside taught me to double-check everything. Turns out that's exactly what AI needs.
Read more →