AI in Clinical Research — “How Can I Help?”

This presentation provides a comprehensive compliance and regulatory framework for the use of artificial intelligence in clinical research, examined from the perspective of study sites and healthcare providers. Using a richly detailed hypothetical, the speakers walk through seven distinct compliance failures that arise when AI tools are deployed across the research enterprise without adequate governance, oversight, and documentation.

The presentation is anchored in the hypothetical by one question: “How can I help?”

This question is framed as the central challenge for health systems, research sites, and providers navigating the rapid integration of AI into clinical trials. The speakers emphasize that AI is not one tool, but an ecosystem deployed at the sponsor level, the site and investigator level, and the regulatory and institutional level, each carrying distinct risks and obligations.

The hypothetical involves five AI tools (e.g., specific to site selection, subject recruitment, informed consent, analytics, and more), each of which generates a compliance issue touching on algorithmic bias, HIPAA violations, IRB review deficiencies, informed consent failures, remote monitoring lapses, and research misconduct detection. Through these scenarios, the presentation maps AI-specific obligations under the full suite of applicable federal regulations.

The presentation concludes with a call to action urging institutions and providers to immediately inventory AI tools used in clinical research, audit vendor Business Associate Agreements for AI-specific gaps, review consent forms for AI disclosure obligations, train IRBs using the updated regulatory framework, implement “human-in-the-loop” oversight workflows, and update research misconduct policies in accordance with ORI requirements.