The Assistant Secretary for Technology Policy (ASTP) is currently moving to overhaul the health IT certification landscape by stripping away dozens of regulatory requirements established over the past several years. This proposal, part of a broader push to reduce regulatory red tape, aims to save the industry an estimated $1.53 billion by simplifying the criteria software developers must meet to stay certified. However, the move has triggered a significant backlash from major medical organizations who argue that less oversight will inevitably lead to a "black box" environment for clinical decision-making tools.

A central point of contention is the removal of "model cards," which function as standardized disclosure documents for AI models by detailing their training data, intended use cases, and known limitations. Medical groups, led by the American Medical Association, contend that these transparency requirements are the only enforceable baselines currently integrated into physician workflows. Without such data, clinicians lose the ability to verify whether an AI tool is safe, effective, or biased before applying its recommendations to patient care.

Beyond AI transparency, the proposed rule would eliminate various privacy and security certification criteria, under the premise that existing health laws already provide sufficient coverage. Provider groups counter that this shift will force hospitals to perform their own technical audits, a complex and expensive task that many are unequipped to handle. They further warn that developers may begin charging for security features that were previously standard, effectively shifting the financial burden of compliance from the technology sector back onto healthcare facilities.

The Assistant Secretary for Technology Policy (ASTP) is currently moving to overhaul the health IT certification landscape by stripping away dozens of regulatory requirements established over the past several years. This proposal, part of a broader push to reduce regulatory red tape, aims to save the industry an estimated $1.53 billion by simplifying the criteria software developers must meet to stay certified. However, the move has triggered a significant backlash from major medical organizations who argue that less oversight will inevitably lead to a "black box" environment for clinical decision-making tools.

A central point of contention is the removal of "model cards," which function as standardized disclosure documents for AI models by detailing their training data, intended use cases, and known limitations. Medical groups, led by the American Medical Association, contend that these transparency requirements are the only enforceable baselines currently integrated into physician workflows. Without such data, clinicians lose the ability to verify whether an AI tool is safe, effective, or biased before applying its recommendations to patient care.

Beyond AI transparency, the proposed rule would eliminate various privacy and security certification criteria, under the premise that existing health laws already provide sufficient coverage. Provider groups counter that this shift will force hospitals to perform their own technical audits, a complex and expensive task that many are unequipped to handle. They further warn that developers may begin charging for security features that were previously standard, effectively shifting the financial burden of compliance from the technology sector back onto healthcare facilities.