Responsible AI Use in Sleep Medicine

In this episode of Talking Sleep, host Dr. Seema Khosla welcomes members of the AASM Artificial Intelligence in Sleep Medicine Committee—Dr. Margarita Oks, Dr. Subaila Zia, Dr. Ramesh Sachdeva, and Matt Anastasi—to discuss their recently published position statement on the responsible use of AI in sleep medicine practices.

Artificial intelligence is rapidly transforming healthcare, from AI-assisted sleep study scoring to clinical documentation tools and insurance claim processing. Yet AI is not a monolith—the technology encompasses various types with different capabilities, risks, and regulatory considerations. Matt Anastasi breaks down the different forms of AI clinicians encounter in practice, while the panel explains what "responsible use" actually means in practical terms.

The updated position statement, notably shorter and more accessible than previous versions, addresses four major pillars: data privacy, fairness and transparency, infrastructure requirements, and medical-legal considerations. The discussion explores critical questions facing sleep medicine practitioners: How do we understand and trust the AI systems we use? What happens when insurance payors deploy AI to deny claims—should we fight AI-generated denials with AI-generated appeals? Do patients need to be informed when AI is used in their care, and how specific must those disclosures be?

The conversation delves into liability concerns that keep practitioners awake at night: If your employer implements AI and it makes an error, who bears responsibility? What about ignoring AI prompts—does that create liability? Dr. Sachdeva explains the concept of vicarious responsibility and how it applies to AI implementation.

Practical implementation guidance is provided through discussion of governance checklists, equity considerations in AI deployment, and the limitations of FDA clearance for AI-assisted sleep study scoring. The experts introduce AASM Link and discuss how practitioners can evaluate AI tools beyond marketing claims, ensuring systems are trained on diverse, representative data sets.

The episode tackles a fundamental question: Is AI use inevitable in sleep medicine, or can practitioners opt out? The panel offers realistic perspectives on integrating AI responsibly while maintaining clinical judgment and patient-centered care.

Whether you're already using AI tools, considering implementation, or resistant to adoption, this episode provides essential guidance on navigating the AI transformation in sleep medicine while upholding professional and ethical standards.

Join us for this timely discussion about balancing innovation with responsibility in the AI era of sleep medicine.