How Emerging Technologies Are Challenging Traditional IRB Oversight
Artificial intelligence (AI), connected wearables, and Software-as-a-Medical-Device (SaMD) are expanding the very definition of a “device.” New technological innovations continue to obscure traditional applications of research tools and interventions, challenging sponsors, CROs, and study monitors to rethink how risk is defined, how submissions are prepared, and how Institutional Review Boards (IRBs) assess safety and ethics.
As FDA guidance evolves for Software as a Medical Device and AI-enabled medical devices, the investigative review process is at a critical moment. While many research leaders are eager for a more streamlined judiciary process for faster approval and use, speed cannot come at the expense of responsibility and clarity. Seeing as the framework for evaluating device risk has outgrown traditional definitions, research stakeholders must adapt now to stay compliant later.
Risk Determination with AI and SaMD
Risk determination sits at the core of IRB review. Traditionally, a study involving a significant risk (SR) device required FDA submission through an Investigational Device Exemption (IDE), while non-significant risk (NSR) studies could proceed with IRB approval and NSR risk device determination.
That distinction is less clear when the “device” is software, data-driven, or adaptive. For example:
- A mobile app that simply collects data may seem low risk — but if it interprets that data to diagnose hypertension, it qualifies as a medical device.
- A wearable monitor may not endanger physical safety, but storing and transmitting sensitive health data introduces security and privacy risks that IRBs must weigh.
- An implantable sensor has direct physiological implications, making mechanical safety, malfunction risk, and data validity equally critical.
When submitting technology-assisted clinical trials for IRB review, sponsors, CROs, and study monitors must err on the side of caution by following up-to-date FDA guidance. Research teams can also lean on pharmacological review frameworks to clarify their technology’s intent, define clinical endpoints, and ground risk assessments in a human-safety-centered rationale.
Redefining What Counts as a Device
The FDA defines a “device” not by its form but by its function and intent. Any technology that diagnoses, cures, mitigates, or prevents disease may be regulated as a device. That can mean modified hardware, apps, or even algorithm-based diagnostics.
This is especially relevant for AI-enabled tools, as the way we teach AI to learn shapes the ethical decisions it makes. AI tends to choose the quickest path to completion, not necessarily the most trustworthy or innovative one. Human investigator-led oversight and decision-making remain indispensable.
Whether your “device” is traditional (mobile or implantable) or non-traditional (AI-enabled software), IRB documentation must always center on human safety. The FDA continues to modify its frameworks for AI-enabled software, assessing not just what a device does to the body but what it does with the data. The risk of an algorithmic error carries the same ethical weight as an incorrect lab test result.
6 Components of Successful IRB Submissions
To streamline IRB review and ensure FDA compliance throughout the trial timeline, investigators, sponsors, CROs, and study monitors should anticipate submissions to address:
1. Comprehensive Context
Explain whether the AI or SaMD poses a potential for physical, data, or algorithmic harm in the context of the proposed study. Detail known and unknown risks to determine whether the study is SR or NSR. Provide background data, prior studies, or validation supporting reliability.
2. Intended Use and Limitations of the Technology
Whether diagnostic, therapeutic, monitoring, or preventive, there must be thorough documentation, including specifications, components, and the scope of embedded functions.
3. Degree of Automation
Research leaders must distinguish trial components that are merely advanced automation from true AI decision-making and data analysis.
4. Change Management Plans
Development of AI-enabled devices may involve adaptive or iterative processes, so updates could alter how data is interpreted or how a subject’s risk is assessed. When algorithmic modifications could influence human safety and data accuracy, research leaders must be prepared to monitor for such changes and submit for IRB approval before deployment. The study protocol should define monitoring procedures and limitations for the adaptations that will be made.
While this end-to-end approach may seem tedious, it protects participant safety and data integrity throughout the trial lifecycle.
5. Consent and Data Security Risks
Participants must explicitly understand how their data will be collected, stored, shared, and potentially used to train AI models. Transparency about data flows — especially where third-party systems or cloud-based analytics are involved — is essential.
Research leaders should make clear verbal and written disclosures about confidentiality risks, and confirm that patients genuinely comprehend the nuances of their participation.
6. Performance and Safety
IRBs need evidence that any AI-enabled technology or SaMD performs as intended and supports the study’s stated hypothesis. Without validation, reviewers cannot confirm alignment with FDA standards.
The higher the quality of information available to review, the more accurately and efficiently the IRB can provide compliant guidance.
With a clear map of the trial thought process, risk determination of the technology involved, and reliable, verifiable data, research leaders can demonstrate competency and IRBs can ensure their trial stays on schedule. Faster, safer clinical care is a shared goal across the board.
Looking Ahead
FDA guidance will continue to evolve as new technological innovations emerge. Investigators, sponsors, and CROs should always consult the most current FDA guidance before submitting IRB documentation.
Ultimately, innovation and ethics must advance together. AI may accelerate clinical discovery and support ethical review, but it cannot take the place of human judgment in clinical evaluation. IRBs across the clinical space safeguard that balance and protect human-centric decision-making above all.
Research teams can navigate this accelerating technological frontier confidently and efficiently by embracing their responsibility of transparency, understanding the nuance of the risks involved in their trial, and aligning with the latest regulatory expectations early and often.
About the Authors:
Raffaella Hart, MS, CIP, serves as Senior Vice President at BRANY, where she leads the organization’s IRB, IBC, and Quality Assurance services. A Certified IRB Professional (CIP) with over 25 years of experience in IBC and IRB administration, she was instrumental in achieving and maintaining BRANY IRB’s full AAHRPP accreditation. Hart holds a Master of Science in Healthcare Science with a focus in Clinical Research Administration from The George Washington University and a B.S. in Psychobiology from Binghamton University. Her work has been featured in IRB Advisor, Clinical Trials, The CenterWatch Monthly, The Monitor, and Seminars in Nuclear Medicine.
Linda Reuter is the Sr. IRB Director at BRANY. Her responsibilities include supervising the IRB staff and maintaining compliance with the IRB Standard Operating Procedures as well as applicable regulations. This includes supervising pre and post IRB meeting activities, expedited reviews, continuing reviews, protocol monitoring, continuing education, development of short and long term strategic goals for the organization, maintaining AAHRPP accreditation standards, policy maintenance and development, and training and mentoring of IRB staff.
Ms. Reuter began her career in IRB Administration at North Shore University Hospital Manhasset, currently part of Northwell Health. She held various positions within the Health System’s IRB program over a 20-year period. Linda formed IRB Consulting, LLC in 2012, providing IRB administrative services, training and education, audit services, and general consulting to numerous IRB programs, including BRANY and HRP Consulting Group, as well as several local institutions.
Ms. Reuter graduated with honors from SUNY Stony Brook with a M.S. in Health Sciences. She holds a B.S. in Biology from SUNY Stony Brook, and an advanced certificate in Health Care Management. She published work related to her Master’s thesis on Lyme disease in the Journal of Immunology, is a contributing author for a book on How to Work With and Within IRBs and has been a certified IRB professional since 2001.
Please let us know if you have a general inquiry or require clinical trial support.
For IRB inquiries, please complete the IRB Contact Form.
