How to Operationalize AI Governance in Clinical Trials

At a recent Society of Quality Assurance (SQA) conference, a particular panel focused on unusual audit questions – a topic that has stayed with me since.

The questions were not unusual at all, but rather a potential signal of where the industry may be headed. Sponsors, auditors, and regulators – all of us – are pushing beyond traditional compliance checklists and into a new era defined by Artificial Intelligence (AI) governance and how AI will be incorporated into clinical trial design and execution.

Whether you are in the role of CEO, Quality Lead, Study Monitor, Biostatistician, or any role involved in running or contributing to clinical trials, these insights are not theoretical. They represent immediate areas of risk and opportunity for implementation.

While AI usage in clinical trials is not new, is the industry truly ready? Do we really understand how to operationalize and apply the concept of AI governance in clinical trials?

How to Operationalize AI Governance

As this is no longer optional, here are my takeaways from the conference, as well as some suggestions:

• Organizations must have documented AI policies, not informal practices
• Sponsors and vendors are expected to provide clear disclosure of AI use
• Auditors may now be required to formally attest whether AI was used in their quality assurance processes

Now is the time to make a fundamental shift. AI is no longer treated as a behind-the-scenes efficiency tool – it is now part of the regulated ecosystem.

The Time is Now – Where to Begin

If AI may influence decisions about data, and processes, it must be governed, transparent, and auditable. While it seems many organizations are behind in formalizing their governance plan, this creates a gap that will increasingly surface during health authority inspections. In other words, we must begin to integrate the principles agreed upon by FDA and EMA regarding AI practices published in January 2026 as an industry.

Begin with the following discussions with your teams:

• How will your organization use AI?
  • For example, will AI influence patient selection, endpoint adjudication and safety signal detection? Think about the context of use and do you need to add a statement of such into the protocol, analysis, and AI validation plan.
• Risk must be considered and evaluated in the context of the AI impact on a clinical trial activity.
  • For example, the risk level of using AI to aid in recruitment optimization may have a light impact whereas using AI to monitor clinical site signals might require a more moderate level of control.
• How about clinical investigators?
  • If AI is used to determine clinical relevance of clinical labs, we should consider how this tool supports eligibility and think about ways to test how clinicians use AI and whether they can override it.
• As a clinical quality professional, will an AI tool inform your risk plan?
  • How will you select which clinical site or participant records to review/audit?

The unifying idea across AI in clinical trials must be treated as a regulated, lifecycle-managed system whose rigor scales with its impact on patient safety and regulatory decisions.

The future of Good Clinical Practice compliance isn’t about answering harder questions during an audit. It’s about embedding proactive practices into daily activities so those questions are easy to answer down the road.

Curious about how AI can support your clinical trial oversight? Begin the conversation now.