Why Read: We have warned our members for years about the dangers of using large language models (LLMs) in life-or-death applications; but this has not prevented many companies from trying to move into medical practice, for example - perhaps an object case for how "LLM-pilled" they are. Now, we are suggesting an immediate, global movement to prevent LLM use in the direct diagnostics or treatment of medical patients.

This may be the shortest, and most important, Global Report I have ever written. - mra

_________

The Ethics of Medicine

The Hippocratic Oath is the oldest and most widely known treatise on medical ethics. It requires new physicians to swear by numerous healing gods and dictates the duties and responsibilities of the physician while treating patients. - doctors.practo.com

With regard to healing the sick, I will devise and order for them the best diet, according to my judgment and means; and I will take care that they suffer no hurt or damage. - English translation from the classic Greek of the Hippocratic Oath; ibid. (See "Upgrades" for full text.)

Primum non nocere is a Latin phrase that means "first, do no harm." - Attributed to 17th-century physician Thomas Sydenham; Wikipedia

A Simple Truth

LLMs cannot be prevented from hallucinating. Therefore, they must not be allowed in any direct human medical diagnostic or treatment.

The good news for doctors

You will not find yourself liable for harm to, or deaths of, patients in your care as a direct result of LLM hallucinations.

The bad news for doctors

You will have to exert much more care regarding the presence of LLMs in any information chain leading to direct patient treatment.

The good news for patients

You will not be inadvertently harmed or killed by the inevitable hallucinations of LLMs introduced into your care program.

The bad news for patients

You will need to be proactively aware of whether you are in any way being diagnosed or treated by LLMs and refuse their use in your health program.

The good news for AI firms and investors

You will no longer have to listen to mountains of hype and untruths about the "safety" of LLMs in direct medical treatment, since such use will be illegal.

The bad news for AI firms and investors

If you have built your dreams and investments on using LLMs for direct patient treatment, you've made a mistake.

The good news for all involved

You won't be responsible for randomly harming or killing people at scale.

Our Credo

(That which we all already know, but which needs to be articulated now with particular urgency):

The medical information supply chain imposes an ethical and legal responsibility.

We therefore propose THE ONE LLM REGULATION THAT THE WORLD MUST ADOPT TODAY:

It shall be against the law, in all nations, to use large language models in the direct diagnostics or treatment of medical patients.

 There. That should do it.

Your comments are always welcome.

Sincerely,

Mark Anderson

[email protected]

 Copyright © 2026 Strategic News Service. Redistribution prohibited without written permission.

DISCLAIMER: NOT INVESTMENT ADVICE

Information and material presented in the SNS Global Report should not be construed as legal, tax, investment, financial, or other advice. Nothing contained in this publication constitutes a solicitation, recommendation, endorsement, or offer by Strategic News Service or any third-party service provider to buy or sell any securities or other financial instruments. This publication is not intended to be a solicitation, offering, or recommendation of any security, commodity, derivative, investment management service, or advisory service and is not commodity trading advice. Strategic News Service does not represent that the securities, products, or services discussed in this publication are suitable or appropriate for any or all investors.

We encourage you to forward your favorite issues of SNS to a friend(s) or colleague(s) 1 time per recipient, provided that you cc [email protected]and that sharing does not result in the publication of the SNS Global Report or its contents in any form except as provided in the SNS Terms of Service (linked below).

To arrange for a speech or consultation by Mark Anderson on subjects in technology and economics, or to schedule a strategic review of your company, email[email protected].

For inquiries about Partnership or Sponsorship Opportunities and/or SNS Events, please contact Berit Anderson, SNS COO, at [email protected].

QUOTES OF THE WEEK 

LLMs and Human Life

From: "Don't trust Generative AI to do your taxes - and don't trust it with people's lives," by Gary Marcus ("Marcus on AI" Substack, 3/5/26):

". . . -- the second part of the title of this essay ("don't trust it with people's lives") shouldbe obvious.

"But, then again, maybe it's not. At least in certain parts of Washington:

"Using a system you shouldn't trust with your taxes to blow people up. 2026 in a nutshell."

We'll note that, while Gary is pointing out the life-and-death responsibility of weapons targeting in the Iran war, it is likely that long-term, far more people may suffer or die as a result of medical LLM hallucinations.

"We're introducing ChatGPT Health, a dedicated experience that securely brings your health information and ChatGPT's intelligence together, to help you feel more informed, prepared, and confident navigating your health." - Opening lines from the January 2026 announcement by OpenAI

No, it does not have intelligence, so it can't use intelligence to navigate your health.

"Claude understands the complexity of medical coding, payer rules, and clinical decisions. [No, it definitely does not understand clinical decisions.] It reviews your records and policies, then shows you where the answers came from. Work that took days now takes minutes." - From Anthropic's new healthcare division webpage

 UPDATES

The Original (Classic) Hippocratic Oath

I swear by Apollo the physician, and Aesculapius the surgeon, likewise Hygeia and Panacea, and call all the gods and goddesses to witness, that I will observe and keep this underwritten oath, to the utmost of my power and judgment.

I will reverence my master who taught me the art. Equally with my parents, will I allow him things necessary for his support, and will consider his sons as brothers. I will teach them my art without reward or agreement; and I will impart all my acquirement, instructions, and whatever I know, to my master's children, as to my own; and likewise to all my pupils, who shall bind and tie themselves by a professional oath, but to none else.

With regard to healing the sick, I will devise and order for them the best diet, according to my judgment and means; and I will take care that they suffer no hurt or damage.

Nor shall any man's entreaty prevail upon me to administer poison to anyone; neither will I counsel any man to do so. Moreover, I will give no sort of medicine to any pregnant woman, with a view to destroy the child.

Further, I will comport myself and use my knowledge in a godly manner.

I will not cut for the stone, but will commit that affair entirely to the surgeons.

Whatsoever house I may enter, my visit shall be for the convenience and advantage of the patient; and I will willingly refrain from doing any injury or wrong from falsehood, and (in an especial manner) from acts of an amorous nature, whatever may be the rank of those who it may be my duty to cure, whether mistress or servant, bond or free.

Whatever, in the course of my practice, I may see or hear (even when not invited), whatever I may happen to obtain knowledge of, if it be not proper to repeat it, I will keep sacred and secret within my own breast.

If I faithfully observe this oath, may I thrive and prosper in my fortune and profession, and live in the estimation of posterity; or on breach thereof, may the reverse be my fate!

SNS Quiz

Q: If IBM's Watson - just a very fast reader with lots of memory - was one of the largest crash-and-burn technical and financial failures in modern history, why would anyone expect LLMs - just a very large predictive typing machine - to be anything more?

A: More money.

Q: If LLMs are to be prevented from direct use in human medicine, does this apply to other AI or ML technologies?

A: No. See the "A Simple Truth" and "Our Credo" above.

Revving Up the BS Factory

Recognized experts from both sides of the "LLMs vs. harm" debate have been hard at work lately, with both OpenAI and Anthropic starting new medical spinouts - and with even Stanford University, my alma mater, finding eloquent ways of evading the fact that LLM hallucinations can kill.

Here's a perfect example of the latter: a February 2025 news release from the Human-Centered Artificial Intelligence (HAI) group at Stanford:

Holistic Evaluation of Large Language Models for Medical Applications

Medical and AI experts build a benchmark for evaluation of LLMs grounded in real-world healthcare needs.

Large language models (LLMs) hold immense potential for improving healthcare, supporting everything from diagnostic decision-making to patient triage. They can now ace standardized medical exams such as the United States Medical Licensing Examination (USMLE). However, evaluating clinical readiness based solely on exam performance is akin to assessing someone's driving ability using only a written test on traffic rules, a recent study finds. 

While LLMs can generate sophisticated responses to healthcare questions, their real-world clinical performance remains under-examined. In fact, a recent JAMA review found that only 5% of evaluations used real patient data and the majority of studies evaluate performance on standardized medical exams. This state of affairs underscores the need for better evaluations that measure performance on real-world medical tasks, preferably using real clinical data when possible.

In 2022, Stanford HAI's Center for Research on Foundation Models developed a benchmarking framework, the Holistic Evaluation of Language Models (HELM), which provides evaluations that are continually updated over time. We leverage the HELM framework for medical applications to create MedHELM. Our team, comprising researchers at the Center for Biomedical Informatics Research (BMIR), Stanford Health Care's Technology and Digital Solutions team (TDS), and Microsoft Health and Life Sciences (HLS), collaborated with clinicians, administrators, and clinical informatics researchers to gather diverse and clinically relevant use cases for LLMs.

SNS members will notice that there is no final column on the far right titled "Number of Harmed or Killed Resulting from LLM Use." But if you want to do an LLM benchmarking study for your PhD, you would think this was a great report -

 ETHERMAIL

Note: Some letters may be republished to include subsequent replies.

SNS: "AGENTS OF CHAOS: How OpenClaw & Tether Accelerate Instability"

[Berit &] Evan,

Mining would draw us farther into space, but discoveries of something really valuable from "space vacuum treatment" would really kick off near-space activity. In the 1950s, with vacuum tube electronics, the Sci Fi guys (almost all men then) wrote about vacuum tubes without tubes. I am hoping along two lines: First, we know the "bad" of prolonged time in space -- radiation exposure, weakening due to lower or no gravity. But one or more of those, especially the lower gravity, might be lifesaving or life-extending for many human conditions. Maybe a giant space hospital or two? 

Similarly, chip firms and others devote extraordinary efforts to having "clean room environments." I am hoping "ordinary space" will provide a faster, cheaper, better version of such an environment. 

I also try not to overlook various forms of entertainment, especially athletics and dance, that might evolve in space. Imagine an NFL-sized audience for 3D football or its equivalent. One Sci Fi story talked about luxury mansions in orbit, where the views and the light to zero gravity led to the creation of space palaces for the ultrarich.

The point being I am hoping for developments on both faster, better, cheaper supply and more variety, quantity and quality in things to demand.

Rollie Cole

Senior Fellow, Sagamore Institute for Policy Research
Author, Wholesale Economic Development
Austin, TX

Rollie,

What a fun vignette! I had not considered whether there might be health benefits to being in space. While there are issues with prolonged space travel, I could see how we might discover that it is actually beneficial for certain debilitating conditions. If we had a place to go up there, people in various limited forms of paralysis might feel a lot freer than down in Earth's gravity...

Evan Anderson

UPCOMING EVENTS 

WHERE'S MARK?

Mark will be speaking at and/or attending: The Nvidia GTC AI Conference & Expo, March 16-19, San Jose Conference Center Life Science Innovation Northwest 2026, April 21-22, Seattle Convention Center * The Future in Review (FiRe) conference, May 31-June 3: Qualcomm Institute, San Diego

In between times, he will be thinking good things about our goat of many years, as she moves on to greener pastures.