Urgent.News

What's breaking now, across thousands of outlets.

AI

Utah expands healthcare AI sandbox with August AI, Nolla Health pilots, adds third-party evaluators

Utah is expanding the state's AI regulatory sandbox program with new pilots to test AI-supported prescription renewals, dermatology and physical therapy, while adding physician oversight and independent third-party evaluators to strengthen safety and accountability.

Utah regulators have announced a plan to permit an artificial intelligence (AI) product to evaluate patients and prescribe new drug treatments without prior human review. Nolla Health, a startup, will launch an AI-powered service in the state that will assess patients for mild-to-moderate acne treatment through a structured intake and the analysis of skin photos.

The AI product may determine if one of a list of low-risk drugs is suitable. During the initial stages of the pilot, the AI's recommendations will be reviewed by a licensed clinician who will sign off on prescriptions. However, Nolla Health may eventually be granted the authority to prescribe drugs without this review. The company raised $4.5 million in funding from investors such as General Catalyst last year, and aims to expand its AI-driven treatment options for various conditions.

This new pilot is part of a major expansion of Utah's "AI sandbox," a highly regarded program that allows the state's Office of Artificial Intelligence Policy (OAIP) to exempt certain regulations for companies testing AI products under state oversight. For more details, continue reading at STAT+.

Written by urgent.news from STAT News's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at fiercehealthcare.com →

More in AI

Debugging a silently failing AI workflow with Agent Core full-link observability

The most difficult AI workflows to debug are the ones that don't crash. A multi-step agent pipeline can quietly return a plausible-looking answer that's actually wrong, without ever throwing an error.

  • AI workflows can quietly produce incorrect answers without errors
  • Multi-step agent pipeline involves fetch, analyze, summarize nodes
  • Instrumentation with OTEL-compatible spans detects faulty node

More from Wednesday 7 October →