Health Law Advisor

The Evolution of Health Care AI Benchmarking

Artificial Intelligence (AI) foundation models have demonstrated impressive performance on medical knowledge tests in recent years, with developers proudly announcing their systems had “passed” or even “outperformed” physicians on standardized medical licensing exams. Headlines touted AI systems achieving scores of 90% or higher on the United States Medical Licensing Examination (USMLE) and similar assessments. However, these multiple-choice evaluations presented a fundamentally misleading picture of AI readiness for health care applications. As we previously noted in our analysis of AI/ML growth in medicine, a significant gap remains between theoretical capabilities demonstrated in controlled environments and practical deployment in clinical settings.

These early benchmarks—predominantly structured as multiple-choice exams or narrow clinical questions—failed to capture how physicians actually practice medicine. Real-world medical practice involves nuanced conversations, contextual decision-making, appropriate hedging in the face of uncertainty, and patient-specific considerations that extend far beyond selecting the correct answer from a predefined list. The gap between benchmark performance and clinical reality remains largely unexamined.

“ERISA, you’ll need a lawyer for that.” Our practice group’s tagline is meant to be a shorthand for the alphabet soup of laws that apply to employee benefits, including the Employee Retirement Income Security Act (ERISA). Employee benefits compliance has many traps for the unwary and is ever evolving. Below, we have provided a primer on current issues of importance in the employee benefits area to help in-house attorneys identify potential risks, mitigate them, and know when to call an outside ERISA lawyer.

1. What Is Old Is New: Get Your Health Plan Governance in Order

Employers that sponsor self-funded health plans have a host of complicated obligations. There are greater potential legal, regulatory, and fiduciary risks than in years past with managing health plans because of increased congressional legislation, increased Department of Labor (DOL) focus on group health plan compliance, and increased group health plan litigation, often by the same plaintiffs’ firms that have been suing 401(k) plans in fee litigation the past 20 years or more.

As the song goes, the Food and Drug Administration’s (“FDA’s”) 2024 Final Rule regulating laboratory-developed tests (“LDTs”) as medical devices (“Final Rule”), is not merely dead—it’s really most sincerely dead.

Perhaps not for good, but for the foreseeable future, at least.

The FDA has let the clock run out on the 60-day time period to appeal the March 31, 2025, decision by the U.S. District Court for the Eastern District of Texas concluding that: 1) the FDA overstepped its authority, and 2) the LDT Final Rule of May 6, 2024, was unlawful. As we explained at that time, the Final Rule would have required virtually all clinical laboratories offering their own LDTs to comply with FDA expectations for medical device manufacturers in phases over a four-year period—with the first compliance deadlines set for May 2025.

The March 2025 opinion by Judge Sean D. Jordan vacated the controversial Final Rule a little more than a month before the first implementation deadlines were to take effect, and remanded the issue back to the FDA.

SB 951, which bolsters existing Oregon law prohibiting the corporate practice of medicine (CPOM), passed the state House of Representatives on May 28 and now awaits the signature of Governor Tina Kotek.

As EBG noted in a recent blog, the majority of states have some form of CPOM restriction. Oregon’s doctrine stretches back to 1947, when the state supreme court in State ex. rel. Sisemore v. Standard Optical Co. of Or. banned corporations from owning medical practices, practicing medicine, or employing physicians.^[1]

Since then, however, Oregon has sought to strengthen its CPOM rules legislatively, as entities have “sought to circumvent the ban through complex ownership structures, contracting practices, and other means,” as SB 951 states.

On May 19, 2025, the U.S. Department of Justice (DOJ) announced a new Civil Rights Fraud Initiative that will leverage the federal False Claims Act (FCA) to investigate and litigate against universities, contractors, health care providers, and other entities that accept federal funds but allegedly violate federal civil rights laws.

The initiative will be led jointly by the DOJ Civil Division’s Fraud Section and the Civil Rights Division—with support from the Criminal Division, federal civil rights agencies, and state partners. 

The initiative implements President Donald Trump’s Executive Order 14173, “Ending Illegal Discrimination and Restoring Merit-Based Opportunity” (January 21, 2025), directing agencies to combat unlawful discrimination through the FCA, and complements Attorney General (AG) Bondi’s February 5 memorandum, “Ending Illegal DEI and DEIA Discrimination and Preferences.”

Those in the tech world and in medicine alike see potential in the use of AI chatbots to support mental health—especially when human support is unavailable, or therapy is unwanted. Others, however, see the risks—especially when chatbots designed for entertainment purposes can disguise themselves as therapists.

So far, some lawmakers agree with the latter. In April, U.S. Senators Peter Welch (D-Vt.) and Alex Padilla (D-Calif.) sent letters to the CEOs of three leading artificial intelligence (AI) chatbot companies asking them to outline, in writing, the steps they are taking to ensure that the human interactions with these AI tools “are not compromising the mental health and safety of minors and their loved ones.”

The concern was real: in October 2024, a Florida parent filed a wrongful death lawsuit in federal district court, alleging that her son committed suicide with a family member’s gun after interacting with an AI chatbot that enabled users to interact with “conversational AI agents, or ‘characters.’” The boy’s mental health allegedly declined to the point where his primary relationships “were with the AI bots which Defendants worked hard to convince him were real people.”

On May 2, 2025, the National Science Foundation (“NSF”) issued a “Policy Notice: Implementation of Standard 15% Indirect Cost Rate” (NSF 25-034) (hereinafter “Policy Notice”) adopting a uniform 15% Indirect Cost Rate (“IDC”) for all new NSF grants and cooperative agreements awarded to Institutions of Higher Education (“IHEs”). The Policy Notice, which became effective May 5, 2025, sets forth a new policy by which NSF will now apply a single, standard IDC “not to exceed 15%” to all future grants and cooperative agreements awarded to IHEs for allowable indirect costs. Currently, IHEs have reported IDCs ranging from 50% to 65%. The Policy Notice allows the awardee organization to “determine the appropriate rate up to this [15%] limit.”

Rationale for the New Policy

Indirect costs, also referred to as “facilities” and “administrative” costs (“F&A”), encompass costs not directly assignable to a specific project or activity but necessary to support the overall research infrastructure of the recipient organization. Historically, awardees seeking to recover indirect costs related to NSF awards have negotiated IDCs on an institution-by-institution basis. These rates were included in Negotiated Indirect Cost Rate Agreements (“NICRAs”), binding upon the institution and the agency, and applied against the Modified Total Direct Costs (“MTDC”) for the project. In contrast to the new uniform 15% rate, NICRAs represent a formally negotiated rate based on an exchange of information with NSF concerning the institution’s general costs and expenditures, including historical cost information, and regularly updated by the institution, often annually. 

On May 12, 2025, the U.S. Department of Justice’s Criminal Division released a new guidance memo on white-collar enforcement priorities in the Trump Administration entitled “Focus, Fairness, and Efficiency in the Fight Against White-Collar Crime.” In this memo, and the accompanying speech by Matthew R. Galeotti, the Trump  Administration’s appointed Head of the Criminal Division, the DOJ reiterated its previously stated commitment to prosecuting illegal immigration, drug cartels, and transnational criminal organizations. For the first time in the new Administration, however, the DOJ clearly articulated new white-collar enforcement priorities, directing Criminal Division white-collar prosecutors to follow three core tenets: focus, fairness, and efficiency. As detailed below, the new memo sets forth the following three priorities:

1. Focus on High-Impact Waste, Fraud, and Abuse Harming Vulnerable Taxpayers

It should be no surprise that the administration is targeting actors that profit through “waste, fraud, and abuse.” The memo sets clear priorities for its prosecutors to investigate, listing as the #1 priority health care fraud and federal program and procurement fraud.  The memo goes on to provide a top 10 list of “high-impact areas”, with “trade and customs fraud, including tariff evasion” as #2.  Heavy focus is given to fraud perpetrated by foreign actors and conduct threatening U.S. national security. Also listed is fraud victimizing U.S. investors, including elder fraud and Ponzi schemes. Appearing as #8 on the list is violations of the Controlled Substances Act and the Federal Food, Drug and Cosmetic Act, including the creation of counterfeit pills laced with fentanyl and the “unlawful distribution of opioids by medical professionals and companies.”

On May 9, 2025, the Departments of Labor, Health and Human Services, and Treasury (collectively, “the Departments”) asked the D.C. federal court to suspend a lawsuit to challenge the legality of the 2024 Rule on the Mental Health Parity and Addiction Equity Act (MHPAEA) while the Departments consider whether to rescind or modify the 2024 Rule.^[1] On May 15, 2025, the Departments released a public statement that they will not enforce the 2024 Final Rule prior to a final decision in the litigation, plus an additional 18 months after the decision.

The public statement on May 15 provides further details regarding the scope of the non-enforcement policy, including clarification that the 2013 MHPAEA rules remain in effect, as does plans’ obligation to develop comparative analyses of non-quantitative treatment limits (“NQTLs”). However, the Departments have not yet provided any indication of the timeline for publishing a notice of proposed rulemaking to rescind or modify the 2024 Rule, and most likely it will take some time for the Departments to determine how exactly the new rule should be designed to better implement the statutory requirements.

[8/28/2025 UPDATE: Following a special session called by Governor Jared Polis, the Colorado legislature passed SB 25B-004 and it was signed by the governor on August 28, 2025. SB 25B-004 will delay the effective date for implementation of SB 24-205, the state’s historic artificial intelligence law, to June 30, 2026, instead of February 1, 2026.]

On May 17, 2024, Colorado Governor Jared Polis signed Colorado’s historic artificial intelligence (AI) consumer protection bill, SB 24-205, colloquially known as “Colorado’s AI Act” (“CAIA”), into law. As we noted at the time, CAIA aims to prevent algorithmic discrimination in AI decision-making that affects “consequential decisions”—including those with a material, legal, or similarly significant effect with respect to health care services and employment decision-making. The bill is scheduled to take effect February 1, 2026.

The same day he signed CAIA, however, Governor Polis addressed a “signing statement” letter to Colorado’s General Assembly articulating his reservations. He urged sponsors, stakeholders, industry leaders, and more to “fine tune” the measure over the next two years to sufficiently protect technology, competition, and innovation in the state.

As the local and national political climate steers toward a less restrictive AI policy, Governor Polis drafted another letter to the Colorado legislature. On May 5, 2025, Polis—along with Attorney General Phil Weiser, Denver Mayor Mike Johnston, and others—requested that CAIA’s effective date be delayed until January 2027.