DISCOVER
The American College of Physicians recently produced a Policy Position paper on artificial intelligence (AI) in healthcare, complete with recommendations and positions. Continue reading to discover our response as well as practical ideas for applying the recommendations.
The American College of Physicians recently produced a Policy Position paper on artificial intelligence (AI) in healthcare, complete with recommendations and positions.
Read the paper
HIA and AI in Healthcare Education
At HIA, we create solutions that allow people to reach their audiences with interactive education, such as Aivio®️, our educational platform loved by healthcare professionals across the US. Supporting the interactive experience of Aivio is our Physician-Controlled AI™️ (PCAI), which answers viewer questions and guides the user through the content their provider or health system has assigned them.
Our partner, WebMD, uses the same technology, branded HealthInteractive™️, to support health systems delivering remote patient education.
Based on our unique and hands-on experience with AI in healthcare, we would like to take this opportunity to respond and provide practical ways to apply the ACP recommendations.
Reference: ACP Policy Position Paper
Artificial Intelligence in the Provision of Health Care: An American College of Physicians Policy Position Paper
Nadia Daneshvar, Deepti Pandita, Shari Erickson, et al. ACP Medical Informatics Committee and the Ethics, Professionalism and Human Rights Committee.
Ann Intern Med. [Epub 4 June 2024]. doi:10.7326/M24-0146
ACP Recommendation & Position Statement #1
AI Should Support, Not Replace Clinical Decision-Making
HIA Response
Decision-making should be supported, not performed by AI. The Aivio doesn’t make any clinical decisions and doesn’t purport to be a clinical decision support (CDS) tool. However, it utilizes AI to help physicians answer patient questions; this is where a clear distinction should be made.
Generated answers left unvalidated by the practicing clinicians are not only insufficient, they’re potentially harmful. Our Physician-Controlled AI™️ (PCAI) is the opposite of Generative AI. Generative AI is trained on unknown sources and creates answers in the moment. PCAI, however, is trained strictly on the answers provided by the doctor/health system. It cannot go rogue and create its answers. All the responses are validated on multiple levels before delivery to the patients.
“The Aivio is designed to extend the reach and impact of practicing clinicians, taking their message to their patients in their own words and keeping them positioned firmly where they belong – as the source of truth for their patients.”
— HIA’s AI Ethics Committee Chair and Chief Medical Officer, Dr. Alidad Ghiassi
Implementation & Practice
AI as a Complement to Clinicians
Implementation: Develop AI systems that provide decision support rather than decision replacement. For example, AI can analyze medical images to highlight areas of concern for a radiologist’s review rather than making final diagnoses. Ensure that AI outputs are easily interpretable and transparent, allowing clinicians to understand the rationale behind AI-generated recommendations.
Practice: Encourage collaboration between AI developers and clinicians to design AI systems that enhance clinical workflows without undermining the clinician's expertise.
ACP Recommendation & Position Statement #2
Aligning AI with Medical Ethics
HIA Response
For too long, technology vendors have created products and then sold them into healthcare with little understanding of the sector. On the contrary, HIA earned its stripes in healthcare.
As a result, we consider nurturing the patient-physician relationship and improving healthcare equity and access for patients as core concepts when developing, testing, and refining our solutions, even those without AI. This is one of the first known solutions that reduces the clinical workload, noise, and misinformation.
Implementation & Practice
Implementation: Develop AI technologies with input from ethical boards and medical professionals to ensure alignment with beneficence, non-maleficence, autonomy, and justice.
Practice: Establish ethical review committees to oversee AI projects and evaluate their impact on patient care and health equity.
ACP Recommendation & Position Statement #3
Transparency in AI Use
HIA Response
While our solutions are not used for direct medical treatment or decision-making, the Aivio is often used to deliver patient education materials, which can be critical to a patient's successful treatment or care plan. Patient communication is the most underleveraged instrument in healthcare.
Implementation & Practice
Implementation: Create guidelines requiring disclosure when AI is involved in patient care.
Practice: Incorporate AI literacy programs and develop consent forms explaining AI’s role.
ACP Recommendation & Position Statement #4
Privacy and Confidentiality
HIA Response
While the ACP position is aimed at products and companies that access, share, store, or exchange patient and clinician data—which HIA doesn’t—we must still side with the ACP here. We don’t collect viewer information, sell data, or condone training AI on patient data without express consent.
Implementation & Practice
Implementation: Adhere to strict data governance, anonymization, and encryption.
Practice: Conduct regular audits and ensure compliance with HIPAA and data protection standards.
ACP Recommendation & Position Statement #5
Clinical Safety and Health Equity
HIA Response
HIA’s human-in-the-loop analytics (AIM™️) keep clinicians informed about patient concerns and needs, improving the quality of care. Controlled AI allows for fine-tuned messaging, including representation in health equity and education.
Implementation & Practice
Implementation: Integrate continuous monitoring and diverse datasets.
Practice: Promote peer-reviewed research and regulatory standards.
ACP Recommendation & Position Statement #6
Reducing Health Disparities
HIA Response
We welcome the ACP’s recommendations and continue to self-regulate while meeting strengthening international AI standards.
Implementation & Practice
Implementation: Ensure representative datasets and bias mitigation.
Practice: Engage in public-private partnerships to establish equitable AI guidelines.
ACP Recommendation & Position Statement #7
Accountability and Governance
HIA Response
We agree oversight is critical but acknowledge unified governance frameworks will take time to mature. The key is values-driven, ethically aligned design from the beginning.
Implementation & Practice
Implementation: Advocate for a unified federal AI strategy.
Practice: Establish clear accountability frameworks for AI developers.
ACP Recommendation & Position Statement #8
Reducing Clinician Burden
HIA Response
HIA aims to reduce the cost and time spent on redundant tasks while simplifying AI setup for content creators. Features like automated Q&A generation and AIM™️ Automatic Intelligence Modification help organizations improve content efficiently.
Implementation & Practice
Implementation: Design AI tools to streamline administrative workflows.
Practice: Pilot AI solutions that improve clinician efficiency and satisfaction.
ACP Recommendation & Position Statement #9
Training for AI-enabled Healthcare
HIA Response
The Aivio has been used for staff training, onboarding, and medical education. One study showed it could excel at empathy training for medical students. We support early and frequent AI education beginning as early as K–12.
Implementation & Practice
Implementation: Integrate AI literacy into medical curricula.
Practice: Partner with educational institutions for ongoing training.
ACP Recommendation & Position Statement #10
Environmental Impact of AI
HIA Response
HIA firmly agrees and looks forward to this research.
Implementation & Practice
Implementation: Study AI’s environmental footprint and reduce energy use.
Practice: Promote green computing and sustainable AI practices.
In Summary
We chose to respond to the ACP recommendations not only to share how we are changing the education environment with our Physician-Controlled AI-supported solutions, but also to expand the conversation by offering practical strategies for implementing the ACP’s guidance.
Our goal is to foster AI that enhances healthcare while maintaining ethical standards and prioritizing patient well-being.
To learn more about HIA, our solutions, and our AI, please visit www.hia.ai
