Summary: Paragon Health Institute’s new report Healthcare AI Regulation: Guidelines for Maintaining Public Protections & Innovation highlights the need for precise, context-specific AI regulation in healthcare. It advocates using existing regulatory agencies, avoiding centralization, and balancing innovation with safety, ensuring AI benefits the healthcare system while fostering competition and protecting public interests.

Key Takeaways:

  • Context-Specific Regulation: AI regulation should define both technology type and healthcare context, ensuring rules address specific risks and use cases.
  • Support for Innovation: The report warns against regulatory centralization, advocating instead for empowering existing agencies to foster innovation while maintaining safety standards.

Paragon Health Institute has released a new report on artificial intelligence in healthcare: Healthcare AI Regulation: Guidelines for Maintaining Public Protections & Innovation. This paper comes at the end of a year when hundreds of AI-related bills have been proposed across state legislatures, and the federal government has been promoting expansive cross-agency AI regulation.

Balancing AI Regulation with Innovation

The central concern of Paragon’s paper is the prevention of AI misregulation that fails to improve public protections while increasing costs and reducing the medical advances policymakers desire most from AI.

The paper illuminates the complexities of AI regulation in healthcare and then proposes guidelines that balance public protections with the need for healthcare innovation. Among the more noteworthy proposals is the recommendation that AI regulation must specify both the type of AI technology and the healthcare context to which each rule applies. In other words, rules would use granular descriptions like “artificial neural networks in medical image analysis” instead of vague language like “healthcare AI.” This emphasis on the combination of technology definition and medical context recognizes that AI risk is closely tied to these two factors.

RELATED: AI’s Impact on HTM

Guidelines for Regulatory Frameworks

The paper also recommends that policymakers use existing regulatory agencies — which already have industry experience — to govern AI in healthcare instead of a centralized AI office or AI czar. The paper argues such centralization would decrease regulatory awareness of industry-specific considerations and risk duplicative rulemaking between the AI office and other government agencies. Instead, the guidelines advocate for regulatory agencies having the personnel, internal AI expertise, and resources necessary to perform their duties properly.

Key Topics Addressed in the Report

The report’s guidelines also touch on matters such as:

  • Aligning regulatory efforts with the FDA’s historic work of evaluating software as a medical device.
  • Handling scenarios where AI-enabled devices have empirically confirmed benefits but whose mechanism of effect is hard to explain.
  • Concerns around algorithmic discrimination and data privacy in AI-enabled systems.

“AI is poised to make remarkable contributions to American healthcare, but these contributions can be jeopardized by a suboptimal regulatory framework,” said Kev Coleman, report author and visiting research fellow. “What we need is an approach that preserves safety standards while discouraging rules that benefit the biggest AI vendors while impeding innovative startups from entering the market.”