A significant challenge for healthcare organizations emerges as nearly nine out of ten leaders report that security or compliance personnel routinely or consistently examine code generated by artificial intelligence before it is deployed, specifically when that code interacts with patient information, as reported by Paubox. Furthermore, Paubox indicates that seven in ten organizations have encountered an AI-developed tool processing such data without the awareness of the IT or security departments.
These automated scripts, integrations, and custom automations are constructed by personnel utilizing AI coding assistants. The inherent limitation, as pointed out, is that a formal code review is restricted solely to the material that has been formally submitted for inspection.
This situation underscores the operational gap in governance, where the pace of AI-assisted development outstrips established oversight mechanisms within the sector.
Why it matters
This highlights a significant governance gap in healthcare infrastructure regarding AI adoption. The reliance on manual code reviews fails to account for unsubmitted or emergent AI-generated tools handling sensitive patient data. Operators must address this blind spot to maintain compliance and security.
