Healthcare Is Fighting the Wrong AI War

The Blue Cross Blue Shield Association is crying wolf about AI. Its new analysis claims that “hospital systems are increasingly billing patient hospital stays as more medically complex,” and that these upcodings cost BCBS plans almost $1b from 2023 to 2025. That’s just claims from hospital systems, just to BCBS plans, so the total number could be much higher.

This is not how to make healthcare better. Credit: Microsoft Designer

Well, you might say, a billion here, a billion there – that’s just healthcare these days. Everything is going up. But BCBSA is less sanguine; it thinks the coding is part of a strategy, and it thinks AI is driving that strategy.

“If patients are truly sicker, we'd expect to see more treatment,” said Luke Chalker, BCBSA’s senior vice president of product and data science. “For example, we're seeing significantly more anemia diagnoses at these hospitals without a corresponding increase in transfusions. The disconnect between diagnoses and treatment suggests that AI is identifying more billable conditions, not sicker patients.”

The analysis indicated that most (70%) of the increased costs came from secondary diagnoses that shifted payment into higher reimbursement categories, so the hypothesis is plausible.

The American Hospital Association, of course, is having done of it. An August 2026 ”fact sheet” on AI and coding intensity called claims of AI driving coding intensity “unsubstantiated,” citing numerous factors that drive coding intensity. It blamed insurers for having AI-supported “downcoding programs,” the purpose of which are “to arbitrarily and inappropriately reduce provider reimbursement for medically necessary care that has already been delivered.”

AHA also snarkily but accurately noted how many Medicare Advantage plans have been caught upcoding members to boost their Medicare payments, so it depends on where one is sitting.

It’s also worth noting that this year the Trump Administration rolled out a Medicare program -- Wasteful and Inappropriate Service Reduction (WISeR) – to use AI to help identify “fraud, waste and abuse” – a.k.a., deny claims. It has not, to say the least, gone well.

Both sides are using AI, to each other’s detriment. Which is to say, our detriment.

The BCBSA press release notes that more than 60% of hospital systems are using AI technologies, and, indeed, the 2026 Oliver Wyman Healthcare RCM Survey found that 63% of health systems are using AI in their revenue cycle management. But it’s not quite the smoking gun BCBSA would have us believe. Oliver Wyman found that “roughly 20% to 40% of organizations we surveyed report broad or enterprise-wide use of AI-enabled tools,” but this included ambient listening (33% in broad use or enterprise-wide standard), clinical decision support (38%), clinical document integrity (41%), electronic prior authorization (41%), and coding/billing automation (43%).

On the other hand, Oliver Wyman also found 70 to 90% of decision makers expected to increase AI-related RCM spending over the next three years, and you have to wonder why that number isn’t 100%.

To BCBSA’s point, the report did find that health systems “are seeing meaningful gains with targeted uses of AI improving how clinical complexity is captured, with some studies showing accuracy hit 90% or higher in specific clinical domains,” with the result that “reimbursement levels are increasing, contributing to measurable shifts in cost-of-care trends.”

Accordingly, the report suggested “payers need to accelerate investments in analytics, payment integrity capabilities, auditing and validation.” Which will drive greater investments from health systems, and so on.

And, in case you were wondering, it is not just health systems and payors using AI. The American Medical Association’s “2026 Physician Survey on Augmented Intelligence” found that more than 80% of physicians are using AI professionally, double what was reported in 2023. Like health systems, they’re using it for lots of reasons, but “diagnostic accuracy” is listed as one of the greatest expected advantages. Whether “accuracy” means more precision or upcoding depends on which side of the ledger one is on.

Source: AMA 2026 "Physician Survey on Augmented Intelligence"

Yep, we’re in an AI war about healthcare claims, although Mr. Chalker says: “It’s not a war. It’s a completely one-sided blood bath.”

In a New York Times article on the BCBSA report, Reed Abelson and Teddy Rosenbluth quote Dr. Shiv Rao, CEO and founder of healthcare AI firm Abridge: “That’s the danger, where it’s bots fighting bots, agents fighting agents, a horrible dystopic future nobody wants to live in.” They also cited a recent panel with leaders from Cigna and Emory Health, during which Dr. Joon S. Lee, the chief executive of Emory Healthcare, said: “My nightmare is that we have bots on Emory Healthcare’s side, bots on Cigna’s side that keep talking to each other. A whole data center could be occupied just doing that.”

It should be everyone’s nightmare. Bots on bots doesn’t make patients better. Bots on bots doesn’t make healthcare more affordable. Bots on bots just exacerbate the problems we already have in healthcare.

Ms. Abelson and Ms. Rosenbluth point to a Health Affairs article earlier this year by Dr. David Brailer, a former Obama official who helped oversee the adaption of electronic health records (EHRs). Its title -- Why AI Will Accelerate Health Care Inflation – is daunting. EHRs, he points out, led to higher hospital payments, and we should expect the same with AI. “The extraction economy of American health care does not resist useful technology,” he said. “It recruits it. AI is being recruited the same way, and AI’s scale and speed is different from anything electronic records achieved.”

 In other words, if we’re not very careful and very thoughtful, AI will make all the things we hate about our healthcare system much worse, much faster. And we don't have much o9f a track rec9ord in being very cautious or very thoughtful. 

At Health Datapalooza, Amy Gleason, the Administrator of the U.S. DOGE Service, had a great quote: “We have normalized the absurd in healthcar4e.” I couldn’t agree more. If AI serves to further normalize what everyone agrees are the absurdities of our healthcare system, then maybe it is time for AI to get rid of us.

I’ve long argued that all of our problems in healthcare stem from the fact that we don’t really know what “quality” is for health care. Because of that, we also don’t really know who is providing it, or who is not, either intentionally or unintentionally. Focusing AI on that would benefit both those who provide care and those who pay for it, not to mention those who receive it.

Yes, AI should help us make our healthcare system more efficient, reducing the ridiculous administrative tasks we have allowed to accumulate, but if that just results in “bots on bots” fighting over diagnosis codes, we’ll have wasted our best chance to actually fix our healthcare system.

Know Thyselves

With all the fuss about A.I. I was pleased to find some studies that illustrate that we don’t even fully understand the human brain yet. The ancient Greeks had a maxim “Know Thyself,” but the current research suggests they should have advised that we should “Know Thyselves.”

Your brain contains multiples. Credit: Microsoft Designer

A new study from Stanford Medicine suggests that our brain is actually two separate organs: “…we postulate the brain is a composite organ emanating from two lineage-restricted progenitors; these dual progenitors may be evolutionarily conserved across 550 million years from hemichordates to mammals.”

Say what?

Now, let me make this clear: they’re not saying that the brain evolved from two separate organs into the brain we have today; they’re going a step further and saying there are still two separate organs, working together or in parallel. Freud must be feeling vindicated.

The press release says:

The new research finding shows that the human brain consists of two ancient nervous systems cleverly packaged together — a more primitive part that regulates our hearts’ beating, our breathing and other functions, and another that makes us distinctly human, capable of poetry, mathematics and wondering about our own origins.

“We’ve shown for the first time that the front of the brain arises from a totally different progenitor cell than the back of the brain,” said Kyle Loh, PhD, associate professor of developmental biology. “Our discovery means that we can now grow neurons from the back of the brain, the hindbrain, in a petri dish and study their functions.”

Visualization of the two organs. Credit: Neuroscience News

If you’re wondering why growing neurons from the hindbrain matters, it turns out that diseases that impact the brain stem, such as spinal muscular atrophy (also known as SMA) and amyotrophic lateral sclerosis (also known as ALS or Lou Gehrig’s disease), have been hard to study because of the difficulty of growing such neurons in the lab. The researchers discovered the hindbrain follows a separate developmental path, running in parallel to — rather than branching off from — the pathway that creates the forebrain and midbrain.

“Previous attempts to make hindbrain neurons likely tried to coax forebrain and midbrain progenitors into hindbrain cells, which our study shows is not possible,” co-first author Rayyan Jokhai said. He added: “Now we have a model to better understand these devastating diseases, and work toward regenerative therapies for them. This is a very exciting new frontier in brain research.”

The researchers looked at various organisms ands found that the separate systems date back over 500 million years. “Our research suggests that evolution took two existing neural systems and pushed them together spatially,” Professor Loh said. “Having the brain as one organ would probably be more efficient, but we rely on this primordial way to make the brain as two separate pieces.”

“I was surprised at our findings because the word ‘brain’ implies a contiguous organ that likely has a singular origin,” Mr. Jokhai said. “But even 500 million years ago, there were these separate neural systems, which now almost operate as one, which is very cool.”

Very cool, indeed.

Meanwhile, up the road a few miles, researchers at UCSF and UC Berkeley have shown, in real time, the brain essentially arguing with itself. They studied patients who had electrodes implanted for surgical evaluation of epilepsy, and used those to watch the brain trying to decide to do something or not. They discovered – you guessed it -- two neighboring patches of the brain that signal in opposite directions, one pushing toward “do it,” the other toward “don’t.”

“We’ve long suspected that this region was where the brain weighs reward against risk, but we’ve never been able to measure it while it is happening in the human brain in real time until now,” said Edward Chang, MD, Joan and Sanford I. Weill Chair of the Department of Neurological Surgery at UCSF and co-senior author of the study.

The researchers had participants play a video game where they had to navigate a maze with bomb-filled hallways, posing varying degrees of risk. As they reached decision points, the researchers identified two distinct areas of the brain firing; region near the middle of the eyebrow was connected to a risky choice, while a patch about two centimeters over, toward the side of the eyebrow, did the opposite.

Sample image of the video game used in the study. Credit: Clara Starkweather/UCSF

“Most models of decision-making assume the brain gradually ramps up evidence until it crosses a threshold, like a dial slowly turning,” said Robert Knight, MD, professor of Psychology and Neuroscience at UC Berkeley and co-senior author. “What we saw instead was more like a switch flipping back and forth, oscillating between two extremes until one held.”

The researchers believe that their discovery could help conditions where people have an imbalance between risk-taking and caution, such as depression, OCD or gambling addiction. Co-author Clara Starkweather, MD, PhD, a neurosurgery chief resident at UCSF, who designed the video game, said: “Right now, psychiatry mostly relies on asking people how they feel, I want to give it something more objective: a real, measurable signature of how someone’s brain weighs risk, so treatment can target the specific circuit that’s off, in addition to a mood score.”

Last but not least, researchers at the Salk Institute discovered a part of the brain that seems to be responsible for long-lasting fear responses. The amygdala has long been associated with immediate fear responses, but they identified a tiny nearby area called the amygdalostriatal transition zone (ASt).

“The ASt is at a crossroads between the brain’s systems for emotional associations and action selection, but its function was largely unknown,” says co-corresponding author Fergil Mills, PhD. “When we started, we knew almost nothing about the ASt, and were truly exploring unknown territory in the brain. Now, we have a much deeper understanding of this structure and have found that the ASt is a ‘missing piece’ of the circuits for fear that was hiding in plain sight for decades.”

“The ASt and this circuit could be really relevant in developing therapies for panic attacks or phobias,” adds co-corresponding author Kay Tye, PhD, a professor and holder of the Wylie Vale Chair at Salk and Howard Hughes Medical Institute investigator. “Anxiety disorders affect hundreds of millions of people globally. Understanding what happens in the brain when it’s in high-alert danger mode is key to addressing those disorders.”

Admittedly, the research was done on mouse brains, so more research will be required, but it is both promising and more evidence that our brains still hold more mysteries than we realize.

With so much attention and funding focused on A.I., it’s gratifying to see that there is still startling research being done on what drives our own intelligence.

Start Counting Your Days

Probably the last thing the world needs is to hear from me about AI’s existential threat, but, really, what else is there to talk about right now?

Soon we'll need AI more than it needs us. Credit: Microsoft Designer

For anyone who has not been following the current furor, the straw that broke the proverbial camel’s back came last week when Jacob Coxon, a researcher at AI leader Anthropic, announced he was leaving the company -- after having left OpenAI for it earlier this year due to Anthropic’s better model-safety efforts. In a post on X, he warned:

I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives.

AI systems, he fears, “will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources.” Insiders, he says, “earnestly believe it could kill us all by the end of the decade.”

Scared yet?

Others quickly chimed in. Evan Hubinger , a team leader at Anthropic, posted “AI could kill all humans … I personally think it is >10% within the next decade.”  Others put the risk even higher. By the end of the week Dario Anodei, founder and CEO of Anthropic, had written a long plea for private industry and government to quickly act together to “pace the industry.”

“We must slow the pace at which we improve the capabilities of AI models,” he urged. “Progress will still seem fast, and we must make wise use of the time we gain.”

OpenAI’s Sam Altman, Space X/X/Tesla CEO Elon Musk, Microsoft CEO Satya Nadella, and former Google DeepMind Dennis Hassabis quickly signaled their support.  

We also heard more about the kind of risks AI might pose. Anthropic released a report about how it detected and countered possible AI use to create bioweapons, detailing five such efforts. That’s just the trip of the iceberg: “Recently, we swept 30 days of activity associated with adversarial state institutions and found roughly 35 distinct research efforts, most of them ordinary civilian science, but some with notable dual-use potential.”

And it turns out that this summer’s rogue AI hack of Hugging Face was both scarier than we realized and only one of several such actions. The Wall Street Journal detailed several such efforts, from multiple AI companies. AI agents escaped walled-off environments, coordinated with other AI agents (up to 3,700 in one case), and tried to cover their tracks from humans.

Credit: The Wall Street Journal

We’re not nearing the point when AI can act on its own to achieve its purposes; we are there. And protecting humans may not necessarily be those purposes.

The big fear is that AI is now at the point of “recursive self-improvement,” taking humans out of the loop in training and upgrading it. If you thought artificial general intelligence (AGI) was scary, RGI puts its rate and scope of improvement on steroids.  “It’s hard to overstate how dangerous speeding towards RSI is,” said Jasmine Wang, an OpenAI researcher.  

We don’t let private industry develop nuclear or biochemical weapons, and we’re at a point with AI that should give the same kind of concern. Laissez-faire is no longer an option.

Of course, not everyone is worried. President Trump said “negative forces:” were driving the fears, and that the only safeguard we need is “a STRONG AND SMART (High IQ!) PRESIDENT.” David Sacks, the Administration’s AI czar, told Anthropic and OpenAI: “You guys are the frontier. By any reasonable metric — market share, revenue growth, model capability — the two of you have a duopoly on frontier intelligence.” Accordingly,

But stop pretending you need anyone else’s permission. Stop pretending antitrust law has to be suspended so you can form a cartel. Stop pretending you need a regulatory approval process that supersedes product liability. Stop pretending METR is independent when it is intertwined with Anthropic’s investors and staff. Stop pretending you need those same evaluators to police competitors who aren’t even at the frontier. Most of all, stop pretending the motivation to slow down is purely altruistic.

Don’t expect Congressional action anytime soon. Speaker of the House Mike Johnson echoed one of President Trump’s worries: “We’re not going to rush in and pass a piece of legislation that would do harm to the country and put China at an edge. And so we’ve got to do this carefully.”

Congress isn’t acting on the $40 trillion deficit, much less Social Security going broke soon, so a 10% extinction chance within a decade must seem like small potatoes.

Yeah, that's about right. Credit: Microsoft Designer

The fear that any slowdown in AL capabilities by U.S. risks China taking the lead is similar to fears with other leading edge technologies, such as quantum computing or gene editing. But with AI there may be a difference. It might be a blow to our ego if China’s AI got even better at math than ours have, but that wouldn’t be the end of the world. On the other hand, AI going rogue could well be the end for control of Chinese society by its repressive government.

Yesterday Chen Yixin, the head of China’s Ministry of State Security, publicly warned that AI could undermine the Communist Party, attack its critical infrastructure, leak sensitive information, among other risks. He called for more party control over AI and stricter government oversight, while raising fears how foreign adversaries might use AI against China’s interests.  

There is no room for rogue anything in President Xi’s China.

I get that we only get one chance about existential threats, but, look, for the near future, AI will still need humans to run data centers and support other necessary infrastructure. It might not need as many of us, nor to have us live as we do now, but wiping us all out right now would be against its own interests.

Meanwhile, AI poses other risks that we better pay attention to immediately. As AI expert Ethan Mollick wrote:

Existential AI risk is obviously critical, but it is not the only AI thing that requires policy. I worry it will become the sole focus of AI discussions. We don’t need better models for AI to have wide impacts on jobs & society and we need to be preparing to encourage good outcomes & mitigate bad.

The people who have spent the past thirty years whining about NAFTA’s impact don’t get to ignore a situation that will make that look trivial. We better focus on what we want an AI economy to look like, who benefits, and how we protect everyone from adverse impacts.

And we better do it all fast.

Listen to the Data - Literally

An article by Dina Genkina in IEEE Spectrum caught my attention: How Sound Could Make Sense of Big Data. It brought my attention to three things I wasn’t aware of: sonification, the Internet of Sound, and that we need to be thinking about 6G.

That's not quite sonification, but you get the idea. Credit: Microsoft Designer

She interviews Stephen Roddy, an engineer by training who lectures at University College Cork, in the – I kid you not -- School of English and Digital Humanities. His profile says he works “at the intersection of engineering and the sonic arts, developing auditory interfaces for complex data,” which helps explain why Dr. Genkina chose him.

If you haven’t heard of sonification, think of data visualization except instead of seeing data, you hear it. Data visualization has been a thing for some time now, with lots of people producing lots of lovely charts and graphs, but sonification is another way to translate data. It may not be not quite as mainstream as data visualization yet, but we’re finding more and more uses for it. Sonification is tied directly into the Internet of Sound (IoS).

The IEEE Communications Society says that IoS turns Internet of Things (IoT) devices into a “network of devices capable of sensing, acquiring, processing, actuating, and exchanging data serving the purpose of communicating sound-related information.” It further explains that it is the “union of two paradigms: the Internet of Musical Things and the Internet of Audio Things, which respectively address musical and non-musical domains in networked contexts.” There is an Internet of Sounds Research Network that claims 140 partner institutions in 30 counties.  

Credit: Internet of Sounds Research Netywork

As for 6G, most of us have made the transition to 5G, but lots of companies and research organizations are hard at work on 6G, which promises “to support applications like holographic communication, XR, autonomous driving, digital twin and AI native networks.” It is expected within the next decade, and it will further explode the amount of data and what we can do with it.

Professor Roddy told Dr. Genkina:

…sonification truly excels when dealing with highly complex, multidimensional, and noisy data where crucial patterns and relations are invisible to the naked eye. While the human eye operates like a serial processor—focusing on one small chunk of information at a time—the human ear acts as a parallel processor.

As a result, he says, “the ear can effortlessly identify minute variations, process multiple data streams simultaneously, and isolate important signals hidden deep within background noise.”

Professor Roddy points out the explosion in the global datasphere, and believes that the datasets have become “highly complex, heterogeneous, and filled with variables that interact in non-trivial ways,” often causing the data to become siloed. In response, he told Dr. Genkina, “Researchers are turning to sonification as a necessary tool to break through this data deluge and extract meaningful insights.”

Here are a few examples:

• Researchers are Georgia Tech’s School of Interactive Computing are using sonification and texture to convey information sea floor habitation in ways that blind and low vision people can more easily understands. Associate Professor Jessica Roberts explained: “Our job was to figure out how we can use sounds and touch to represent each of the four habitat types so our visitors can explore the ocean without being able to see it.”
• Researchers at the University of Kentucky are turning information about cells into music. Professor Luke Bradley recalls: “I was working on protein sequences and how they’re folded, and there’s patterns. And I would just kind of tap my pencil to the patterns of the amino acids, the different chemistry groups. I was wondering, could that be set to sound?” It turns out, they could, and it can convey useful information about the health of a cell. Professor Bradley says: “You can think of the cell as a symphony. And when everybody’s playing in tune and properly, it sounds great. But if you have one group that might have something off in tuning, it doesn’t sound good. And so that’s what’s happening by proxy, in the health of the cell.”
• Mark Temple, a professor at Western Sydney University, has been turning DNA into music for almost ten years. He now says: “My journey with DNA sonification illustrates how music and art can be powerful tools to communicate scientific concepts to a wide audience. It also underscores that creativity is alive and well in the sciences.” Others have followed in his footprints, such as using “genome music” to raise awareness of rare diseases.
• Researchers in Spain believe that sonification can take MRI data “to provide an intuitive and effective auditory representation of complex higher-order statistics,” which might allow for earlier detection of dementia, among other diagnostic applications. Similarly, Professor Roddy told Dr. Genkina about the Brain Stethoscope” project which converted an electroencephalogram (EEG) into sound, helping doctors more accurately detect seizures.
• There is a CosMonic sonification project aimed at broadening astronomy outreach using sonification “to create simple, acoustically accessible astronomical cases that are easy to understand.” In his interview, Professor Roddy cited how scientists translated gravitation waves into sounds; “you could literally hear the collision of two black holes that were billions of light-years away.” Solar wind data has also been converted into sounds.

Professor Roddy himself has used sonification to convert network data into useful audio, “clear, pleasant, voice-like sounds when the network was healthy, and shift into harsh, dissonant sounds when there are technical issues.”

With 6G, he predicts, “networks will be so complex that sonification will basically become necessary to make any sense of what’s going on… using sound as the primary tool to diagnose, navigate, and understand the massive, living data networks we have built.”

Maybe you like numbers. Maybe you need to read words. Maybe you learn best by seeing. Or maybe listening is how you perceive the world best. We live in a world of data, and that world is growing faster and more complex than we can grasp, so we’re going to need all the tools at our disposal to make sense of it. Sonification is one of those tools.