Trump administration's GenAI Medicare reviews raise legal concerns

As we start the new year, I am compelled to share my thoughts on some perplexing news that fits into the intersection of logical reasoning located at the convergence of three columns I wrote for the Daily Journal last year. It's news that will impact us all, sooner or later, in 2026 and beyond, as we age and need to access and utilize our Medicare benefits to fund our medical expenses.

One of my columns focused on critiquing the expanding--and almost ubiquitous, and therefore mandated--use of legal technology, especially generative AI (GenAI), in the practice of law, now and going forward. I explored why lawyers and courts must not only learn how to use these not-so-new tools of our trade, but also how to command them.  I also provided some lessons learned on how we can--and must--competently navigate GenAI and cybersecurity in order to avoid being sanctioned by our regulatory bodies for their unethical or careless use. (See "How lawyers can navigate AI and cybersecurity," published Nov. 24, 2025, at https://www.dailyjournal.com/articles/388737-how-lawyers-can-navigate-ai-and-cybersecurity.)

Another column discussed the purpose and function of insurance, which I wrote in the wake of the Eaton Fire, as we, as a society, face the growing risk of financial uncertainty amid the ever-increasing frequency of natural and man-made disasters, whether weather- and economically related. One example is Congress's refusal to extend the tax subsidies of the Affordable Care Act that expired at midnight on Dec. 31, 2025.  A decision that I believe was foolish and will end up driving millions of Americans and their families, along with some health care providers and hospitals, into bankruptcy court. (See "Insurance survival guide for modern disasters," published Jan. 24, 2025, at https://www.dailyjournal.com/articles/383013-insurance-survival-guide-for-modern-disasters.) 

While the third column in this constellation of thought discusses how the Heritage Foundation's 2024 manifesto, titled 2025 Mandate for Leadership posits that "all federal executive power vest solely in the person of the President and accordingly, 'it is the President's agenda that should matter to the [executive branch] departments and agencies [i.e., the Federal Reserve], not their own'."  Further, the 2025 Mandate suggested at the time "that the next conservative Administration [now the current Trump administration] should recognize the constitutional obligation of the executive branch to restrain the excesses of both the legislative and judicial branches, emphasizing the necessity of inter-branch pushback as a positive aspect."  (See "The Project 2025 Mandate: Implications for the Administrative State and the U.S. Constitution," published July 19, 2024, at https://www.dailyjournal.com/articles/379782-the-project-2025-mandate-implications-for-the-administrative-state-and-the-u-s-constitution.)

The point of convergence of my three columns manifests itself today with the fact that the Trump administration has not only succeeded in the outright emasculation of the federal legislative branch, but also in bypassing the still-functioning trial and appellate federal courts to gain access to a Supreme Court that uses its "shadow docket" to almost uniformly approve what many legal scholars have determined to be the administration's questionable, if not downright illegal, actions. 

Additionally, Trump is using executive orders to lay the groundwork for assuring that the "tech bros" who supported his reelection and are now in charge of creating new and innovative GenAI systems that already incorporate the unbridled subroutines Trump supports--and that are already infected with the anti-woke, inherent racial and gender bias of their developers and our society as a whole--to create the LLMs used to train them. (See Trump's Executive Order 14319, dated July 23, 2025, at https://www.federalregister.gov/documents/2025/07/28/2025-14217/preventing-woke-ai-in-the-federal-government.) As such, these technology titans have also been charged by the administration to only create systems that reinforce these preexisting racial and gender biases.  Further, Trump has just signed another executive order that attempts to hamstring the implementation of any state's regulation of GenAI that are already being sold to our government. (See "Ensuring a National Policy Framework for Artificial Intelligence," dated Dec. 11, 2025, at https://www.whitehouse.gov/presidential-actions/2025/12/eliminating-state-law-obstruction-of-national-artificial-intelligence-policy/.) As a result, these new anti-woke GenAIs have already begun to replace the scores of fired and otherwise terminated federal employees and will be used to control the tax-payer-financed, heretofore human-controlled processes that make--or will make--decisions on just about every critical function of our government, including who will receive their Social Security payments. (See Social Security Administration 'In Turmoil' as New Reporting Details Damage Done by Trump Cuts, dated Dec. 30, 2025, at https://www.commondreams.org/news/trump-social-security-administration-2674837440.) And starting on Jan. 1, 2026, some 6.4 million Americans already enrolled in traditional Medicare and residing in Arizona, New Jersey, Ohio, Oklahoma, Texas and Washington state will be drafted into and become part of an already funded pilot program using GenAI, known as the Wasteful and Inappropriate Service Reduction (WISeR) model. Under this program, WISeR will be required to determine if 17 different medical treatments prescribed by their physicians are really needed by the senior patients--and will thus be paid for by Medicare from the funds these seniors see taken out of their monthly Social Security checks and used to underwrite the Medicare risk pool. (See WISeR (Wasteful and Inappropriate Service Reduction) Model, at https://www.cms.gov/priorities/innovation/innovation-models/wiser.)

According to information posted on the U.S. Centers for Medicare & Medicaid Services website (CMS.gov), which is managed and paid for by CMS, the agency is currently led by Dr. Mehmet Oz, who reports to Health and Human Services Secretary Robert Kennedy, Jr. CMS is the federal agency that provides health coverage to more than 160 million Americans through Medicare, Medicaid, the Children's Health Insurance Program and the Health Insurance Marketplace.

So, I have a few questions:  Who are the WISeR participants? How were they selected? How will they protect our private personal medical and financial information? How will they be monitored and controlled? What's inside their models? How do the models work?  Who will be responsible for the models' misdiagnosis?

Here's some of what I was able to find out about WISeR:

WISeR model participants

One final question: what's in it for these participants? Again, as stated on the WISeR website, the model participants will receive a percentage of the savings associated with averted wasteful, inappropriate care as a result of their reviews. That percentage will be adjusted based on the participant's performance on measures related to the process, including provider experience.

One of my biggest concerns about the unfettered use of GenAI in the public sphere is the same one I've had for nearly 30 years. I first expressed it way back when I was appointed as a public member to the National Association of Insurance Commissioners to represent the interests of California insurance consumers, when the leading property and casualty insurers sought the body's approval to introduce models guided by propriety software to predict and price earthquake, windstorm, firestorm and other catastrophic-based insurance policies.  At that time, I stated:

"Insurance Regulators should not sanction the use of proprietary earthquake catastrophe models in the regulatory process as long as these models are unverifiable "black boxes." If regulators rely on such unverifiable models, they will be creating several public policy risks (1) they will be abrogating their responsibility for determining the fairness and equity of rate filings, (2) they will not be in compliance with the principles of ratemaking developed by the Causality Actuarial Society, (3) they will have illegally transferred their responsibility to the hands of unregulated groups with financial interests without appropriate checks and balances to protect the public interest, (4) they will have developed unsound public policy. To mitigate these risks, this paper offers a list of guidelines for the permissible use of earthquake catastrophe models. To set the state for these guidelines, this paper identifies the actuarial and social implications of injecting these technologies into the regulatory landscape."

And it is still my informed opinion today.  You can download and read the entirety of my critique of non-publicly vetted and controlled computer algorithms used to make public policy decisions that are shieled by "black boxes" in Risky Business: Proprietary Modeling and Insurance Ratemaking, published as part of the NAIC Symposium on Catastrophe Modeling in the Spring 1997 (March 1997) edition of the Journal of Insurance Regulation, at https://naic.soutronglobal.net/Portal/Public/en-US/DownloadImageFile.ashx?objectId=9370&ownerType=0&ownerId=25826

As I shared with some of my fellow seasoned female lawyers as we prepared for the new year, quoting the late great actor Betty Davis' immortal character Margo Channing in the classic movie All About Eve, please "fasten your seatbelts, it's going to be a bumpy ... [next 3 years]."