ADLM Calls for CLIA Updates to Support Safe AI Use in Laboratory Medicine

Learned models can fail in case-specific ways or change behavior after updates, making errors harder to detect, while generative systems may produce inaccurate or incomplete information that compromises patient safety. To address these risks, the Association for Diagnostics & Laboratory Medicine (ADLM) has outlined targeted CLIA updates to support safe, reliable AI use in laboratory medicine.

ADLM submitted comments on September 15, 2026, in response to a federal request for information on potential CLIA updates. The association recommends that emerging AI tools remain subject to the same professional expertise, quality systems, validation requirements, and monitoring standards that govern traditional clinical testing. The organization characterizes these as targeted modifications within CLIA’s existing framework to maintain accurate and reliable results.

ADLM notes that traditional software has long supported high-quality testing by verifying, interpreting, and reporting results, but artificial intelligence and machine learning are increasingly integral to these activities. The association underscores that errors in conventional software generally affect every case meeting the same programmed conditions, whereas AI models can yield case-specific errors that are harder to troubleshoot. It also cautions that generative AI may produce inaccurate or unsupported information, omit clinically important facts, or change behavior following updates to the model, prompt, or knowledge base.

ADLM urges the Centers for Medicare & Medicaid Services and the Centers for Disease Control and Prevention to establish federal oversight of AI-based tools that avoids unnecessary duplication. Under this approach, the Food and Drug Administration (FDA) would regulate specific software products or medical devices, while CLIA would govern laboratories’ responsibility for accurate and reliable testing. 

The association recommends that CLIA distinguish conventional software from learned models and that laboratories reflect these differences in validation and performance monitoring. It also supports a risk-based, technology-appropriate approach that establishes clear quality expectations while allowing laboratory directors and qualified professionals to determine scientifically appropriate methods for meeting them.

This emphasis on laboratory responsibility extends to facilities that independently analyze patient-specific laboratory data or provide specialized interpretation that generates or contributes to a result or interpretation for clinical use. ADLM recommends treating these activities as part of the total testing process, subject to appropriate CLIA oversight. Its comments therefore frame AI-based tools as components of that process rather than standalone software requiring a separate regulatory structure, reinforcing the role of laboratory oversight as AI use expands.

ADLM also advises that when a facility independently analyzes patient-specific laboratory data or provides specialized interpretation that generates or helps generate a test result or interpretation for clinical use, those activities should be considered part of the total testing process and come under appropriate CLIA oversight. 

Overall, the comments emphasize assessing AI-based tools as elements of the total testing process rather than as standalone software requiring a separate regulatory structure. The organization positions continued laboratory oversight as essential as AI becomes more prevalent in laboratory medicine.

“AI has the potential to support tremendous advances in laboratory medicine, but innovation in this area must be balanced with the need to ensure test quality and patient safety. Laboratories have the expertise and quality systems needed to evaluate, implement, and continuously monitor AI tools, making continued laboratory oversight essential to the responsible use of AI in clinical laboratory testing,” said ADLM President Dr. Stanley F. Lo.

Related Links
ADLM