Marin Voice: AI can help a health system being asked to do more with less
Not long ago, it took a month and countless doctors, blood draws and electrocardiograms, plus scouring the internet for a cardiologist to double-check my chart and catch a missed mycoplasma pneumonia diagnosis. While I was relieved to finally know why my sickness had been prolonged, I couldn’t help but wonder how it was that so many highly trained professionals could overlook something so obvious.
It really felt like I was working against a system that was supposed to be helping me.
What I missed at the moment was that this misdiagnosis was not a reflection of the doctors, but rather a national systemic issue within the health care industry. Clinicians are being asked to do more with fewer resources, less time and a decreased budget. Patients are feeling the effect.
In March 2025, in a video announcing the reorganization of the Department of Health and Human Resources, Secretary Robert F. Kennedy Jr. promised that the agency would do “more with less.” For patients, “less” appears to be shorter appointments, fragmented care and providers exhausted with case overload.
My experience is not a unique one. Many Americans are experiencing misdiagnoses in the height of recent defunding, carried out by the Trump administration. For me, my issues cost me a month; for others, it costs them their stability, sanity and sometimes even their life.
At the beginning of 2025, 6-year-old Daniel Witten was hospitalized after developing serious complications that turned out to be a cavernous malformation, despite doctors originally diagnosing him with the flu. While doctors said there was little they could do, his mother researched possible treatments on her phone. She eventually found a neurosurgeon specializing in cavernoma who was able to successfully operate, saving her son. Imagine the pain that Witten could have avoided if he and his mother had known of a streamlined artificial-intelligence software that could reveal his true diagnosis.
There are others not so lucky. According to a study published in July of 2023 by the BMJ journals, 371,000 people die every year from a misdiagnosis, like college student Sam Terblanche, who died of sepsis, after being discharged with “acute viral syndrome.”
It is absurd that although the U.S has the highest gross domestic product in the world, yet it seemingly refuses to invest in universal health care, unlike the 72 other countries around the world that do.
As of 2025, the average cost of a family health plan is nearly $27,000. Considering that the U.S. poverty line for a family of four is $32,150, health care has not only failed through its fragmented care, but it has become outright unaffordable. Considering the defunding of essential health-insurance programs, the need for affordable care has only increased. AI is not a cure at all, but in a system stretched beyond its capacity, it may be one of the few scalable tools clinicians have left.
AI systems are being trained on large clinical datasets and are already demonstrating an ability to flag overlooked conditions, recommend streamlined treatment plans, and prioritize urgent cases. While the current software isn’t replacing physicians yet, it is providing services that can act as a second set of eyes to a time-limited care system.
Concerns around AI medical software are understandable. I fear for the erosion of patient privacy, mass unemployment and the true accuracy of these systems.
But the health care system is suffering, and it’s not front-page leaders advocating for defunding who are facing the brunt of it; it’s everyday American citizens who experience an expensive, rushed, dismissive and unhelpful system. So, if AI offers a tangible and implementable solution, we must engage, learn and improve it before it can be used for harm.
Ella Rubenstein, of Ross, is a student at The Branson School.