Policy Playbook: Executive Order on Artificial Intelligence (AI)

Author: Phillip Groden, MD (PGY-3, Mount Sinai Emergency Medicine) // Reviewed by: Summer Chavez, DO, MPH, MPM (Attending Physician, University of Houston); Brit Long, MD (@long_brit); Alex Koyfman, MD (@EMHighAK)

Policy Playbook returns to emDOCs with a concise summary of the latest developments in emergency medicine-related health policy over the summer months. In today’s post, we’ll highlight the most important aspects of each topic.

What’s the issue?

Collins Dictionary has dubbed “AI” as our Word of the Year 2023, which it defines as “the modeling of human mental functions by computer programs”.¹ And apparently our linguist experts are not the only ones to notice the staying power of this novel technology. Late last month, the Biden Administration issued an extensive Executive Order aimed at establishing guidelines for the safe, effective utilization and proliferation of AI technologies throughout the federal government, private sector, and international community.²

With recommendations spread across numerous domains— from Promoting Innovation & Competition to Advancing Equity & Civil Rights— the Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence outlines numerous targeted and diverse recommendations.² The Order comes on the back of another piece of related White House recommendations released one year ago titled the Blueprint for an AI Bill of Rights, which outlines general principles aimed at protecting American citizens from automated systems with the potential to impact “rights, opportunities, or access to critical resources or services”. The Blueprint is additionally accompanied by a pragmatic handbook, From Principles to Practice, for incorporating these guidelines into real-world applications.³

These recent actions by the Biden administration have been largely applauded by AI stakeholders— many of whom took to calling for increased regulation in emergent “black box” AI technologies earlier this year. The Executive Order also aligns the U.S with other Western powers that have outpaced the U.S in developing AI regulations, such as the European Union’s Artificial Intelligence Act and Australia’s Responsible AI Network.^4,5  A majority of  voters support the President’s latest actions on AI, with 69% of voters across the political spectrum indicating their approval of the document.⁶  Within the Executive Order are numerous targeted recommendations relevant to healthcare professionals and to the health of our nation.

Why does this matter?

AI technology has been emerging in the health sciences sector and in patient care applications for years. AI-assisted image processing software has been trained to detect the presence of pathology across a variety of modalities with startling accuracy. Deep learning models on wearable devices are capable of sensing aberrations in heart rates and alerting wearers to the presence of dangerous arrhythmias. And natural language processing models are able to digest large amounts of medical record text and abstract useful data for health systems or insurance companies, to list a few current applications. While the extent of its future footprint on patient care and the workforce are up for debate, it’s clear that AI in healthcare is here to stay.⁷

In a recent survey of over 1,000 physicians across dozens of specialties, a majority (65%) indicated they were “somewhat” to “very” concerned about future AI technologies influencing diagnosis and treatment decisions.⁸Rumblings in medical school threads are ripe with apprehensions over which prospective specialties to avoid over concerns of job security due to AI intrusion. Experts and physicians alike have also raised concerns about the potential for emerging AI technologies to actually harm patients, from providing convincing misinformation to patients via “chatbots” to falsely interpreting patient data for decision makers.⁹

The Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence outlines numerous recommendations related to the healthcare and labor sectors in addition to the health and security of our society at large: ²

• Mobilize the Defense Production Act to require developers to share critical safety information on foundational AI models with the potential for risks to our national security, economy, or public health.
• Develop tools and regulations via the National Institute of Standards and Technology to ensure the safety of AI products before public release.
• Establish standards for biologic synthesis screening as a criteria for federal funding for any agency that finances life sciences research to prevent the use of AI for the development of dangerous biologics.
• Direct the Department of Justice to formulate best practices for investigating and prosecuting civil rights violations within AI algorithms.
• Promote the responsible utilization of AI within the healthcare field for the benefit of our patients, including the creation of critical, novel therapeutics.
• Develop a safety program within the Department of Health and Human Services to collect and remedy reports of unsafe utilization of AI in the healthcare field.
• Outline principles to maximize the benefits of AI on the workforce while minimizing harms in relation to job displacement and health, safety, and equity concerns.
• Ensure the rapid implementation of AI standards with international partners and mobilize the utilization of AI within the international community to solve global challenges in a manner that is safe and equitable.

While likely not sufficient to assuage all fears about the future of AI in healthcare and society, these recommendations should provide some solace that the federal government is being thoughtful in their approach to maximizing the upside of this emerging technology while attempting to mitigate its very plausible harms for the benefit of our patients, the healthcare workforce, and the nation.

What can I do about it?

• Review the full Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence here.
• Read the Blueprint for an AI Bill of Rights and the From Principles to Practice guide here.
• Learn more about AI applications in healthcare (and earn CME credit!) here.

This post is a collaboration between emDocs and the EMRA Health Policy Committee.

References

1. American Medical Association. (2023). “Inside the proposed 2024 Medicare physician pay schedule.” from https://www.ama-assn.org/practice-management/medicare-medicaid/inside-proposed-2024-medicare-physician-pay-schedule.
2. Seidenwurm, D. J. and J. H. Burleson (2014). “The medicare conversion factor.” AJNR Am J Neuroradiol 35(2): 242-243. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7965749/#:~:text=It%20is%20calculated%20by%20use,the%20regulations%20governing%20covered%20services.
3. American Medical Assoication. (2023). “Inside the proposed 2024 Medicare physician pay schedule.” from https://www.ama-assn.org/practice-management/medicare-medicaid/inside-proposed-2024-medicare-physician-pay-schedule.
4. Rosso, R. J. (2023). Medicare and Budget Sequestration, Congressional Research Service. https://crsreports.congress.gov/product/pdf/R/R45106
5. American Medical Association. (2023). “Medicare physician pay must track inflation—every year.” from https://www.ama-assn.org/practice-management/medicare-medicaid/medicare-physician-pay-must-track-inflation-every-year.
6. Khullar, D., et al. (2021). “Time and Financial Costs for Physician Practices to Participate in the Medicare Merit-based Incentive Payment System: A Qualitative Study.” JAMA Health Forum 2(5): e210527. https://jamanetwork.com/journals/jama-health-forum/fullarticle/2779947
7. Cosgrove, J. (2018). Small and Rural Practices’ Experiences in Previous Programs and Expected Performance in the Merit-based Incentive Payment System, Government Accountability Office. https://www.gao.gov/assets/gao-18-428.pdf
8. Bond, A. M., et al. (2022). “Association Between Individual Primary Care Physician Merit-based Incentive Payment System Score and Measures of Process and Patient Outcomes.” JAMA 328(21): 2136-2146. https://jamanetwork.com/journals/jama/article-abstract/2799153
9. Center for Medicare and Medicaid Services. (2023). Fact Sheet: Key Information on the Process for the First Round of Negotiations for the Medicare Drug Price Negotiation Program. https://www.cms.gov/files/document/fact-sheet-negotiation-process-flow.pdf
10. Kaiser Family Foundation. (2023). “FAQs about the Inflation Reduction Act’s Medicare Drug Price Negotiation Program.” from https://www.kff.org/medicare/issue-brief/faqs-about-the-inflation-reduction-acts-medicare-drug-price-negotiation-program/.