The guidance is a tech-positive approach that enables AI to support a digital-first future, said Kyle Zebley, ATA CEO and executive director of ATA Action.
The American Telemedicine Association has updated its Policy Principles on Artificial Intelligence in a framework that the ATA said promotes the safe, ethical and scalable deployment of AI in healthcare.
The ATA originally published guidelines in October 2023. Since then, the ATA said its member-led AI Work Group has added guidance on self-regulatory practices, ongoing performance monitoring, continuous improvement and data security to its Policy Principles on Artificial Intelligence.
Key components include accountability and engagement, transparency and explainability, safeguards to mitigate bias, clear regulatory guardrails, validation and performance monitoring, privacy and data security, and economic and workforce evolution.
WHY THIS MATTERS
There is no single federal law for AI compliance. This is shifting toward a more centralized federal approach, but current regulations are based on a patchwork of state and federal frameworks.
“These principles reflect a commitment to thoughtful governance, responsible policymaking, and a tech-positive approach that empowers providers, protects patients, and enables the full potential of AI to support a digital-first future of care,” said Kyle Zebley, ATA CEO and executive director of ATA Action.
The updated guidance states there should be clear roles and responsibilities among developers, infrastructure providers and clinical deployers.
Developers and deployers should provide transparency about the AI system’s intended purpose and the nature of the patient information it analyzes. They should also lead in establishing self-regulatory practices and in providing transparency regarding what the health data AI systems collect and process, the ATA said.
Regulatory frameworks should protect proprietary information to support innovation and investment.
AI that directly interacts with users should clearly disclose to users that they are engaging in an AI interaction. When AI is used to convey clinical information or influence patient care, that use should also be made explicit to the patient.
To guard against consent fatigue, disclosure requirements should avoid unnecessary repetition.
AI used in clinical care should be routinely evaluated for bias and its impact on health access, with the recognition that fairness standards may shift over time.
Technology should be designed and deployed to support healthcare providers in delivering high-quality care, reduce administrative burden and help address workforce shortages, the ATA said.
THE LARGER TREND
Unified, harmonized and risk-based AI laws and regulatory frameworks in combination with self-regulatory industry practices are essential for both the consistent compliance across the nation by providers, developers and researchers and the efficient adoption of trusted, validated AI systems and tools, the ATA said.
ON THE RECORD
“The ATA and ATA Action are eager to continue to partner with the Trump Administration, our bipartisan champions in Congress, and state policymakers to ensure AI in healthcare policy strikes the right balance, one that promotes innovation, provides regulatory clarity, and keeps patients firmly at the center of healthcare decision-making,” Zebley said.
Aaron T. Maguregui, chair of the ATA AI Work Group, and partner and digital health attorney with Foley & Lardner, said, “This is a tech-positive, workable framework built by practitioners actively deploying and governing AI in healthcare, and the ATA is excited to help shape a future where AI strengthens care, expands access and earns lasting trust.”
Email the writer: [email protected]
link