The HTI-1 Final Rule sets first-of-its-kind transparency requirements for AI and predictive algorithms used in certified health IT, aiming to make clinical users aware of the algorithms supporting their decision-making. It also includes the adoption of the United States Core Data for Interoperability (USCDI) Version 3, effective from January 2026, focusing on more accurate patient data to promote equity and reduce disparities.

Additionally, the rule revises definitions and exceptions around information blocking to promote more efficient and secure information exchange. Health IT developers are required to comply with new conditions and reporting metrics, such as the Insights Condition, which mandates reporting on interoperability and transparency metrics.


Vendors need to adapt to new standards and criteria, ensure their products meet transparency and interoperability requirements, and prepare for regular reporting as mandated by the Insights Condition. Providers should be aware of their roles under the information blocking regulations and consider how AI technologies they use are certified under the new ONC rules. They must also navigate the updated exceptions to ensure compliance while optimizing the use of AI in their practices.

Looking ahead, the ONC has indicated that further updates (HTI-2) are planned, suggesting a continuous evolution of regulations in this space. Providers and vendors should stay informed about these developments to ensure they remain compliant and effectively leverage AI technologies to improve healthcare delivery.

HTI-1 needs more teeth
While the HTI-1 Final Rule is a step in the right direction, it lacks the necessary "teeth" to be truly effective. The rule requires clear definitions and guidelines for algorithm-based and model-based predictive DSIs, user and auditor access to information about the DSIs' design, development, training, and evaluation processes, implementation of risk management practices, and mandatory public disclosure of summary information regarding the risk management practices applied to predictive DSIs.

However, the rule falls short in addressing the challenges posed by "black box" LLMs that do not disclose source attributes. These models, which are increasingly being used in healthcare AI applications, lack transparency and may perpetuate biases or inequities. To ensure the responsible development and deployment of AI in healthcare, the ONC must establish more stringent requirements for the disclosure of source attributes and the evaluation of model fairness.

U.S. Healthcare Homepage