By Roshan Sapkota
Junior year of college, my roommates and I were taking computer science classes when ChatGPT dropped. We huddled around a laptop, watching it reason through problems none of us expected a machine to touch. I was pre-med and already thinking about what this could mean for medicine, though I didn’t yet have the clinical experience to know where.
Medical school gave me that experience. In the clinic, I noticed how much of the visit was consumed collecting information for administrative, billing, and reimbursement purposes rather than clinical insight. Physicians told me the same thing: this data matters, but it could be collected beforehand. My co-founders and I saw a clear application for AI: offload that groundwork before the visit. Patients loved the idea. Physicians said it was worth building. What followed was eighteen months of education about the legislative and structural failures that keep homegrown solutions from reaching patients. I want to briefly share three of them.
The Cost of Entry Is Prohibitively High. A single EHR integration can cost $30,000 to $150,0001. HIPAA compliance adds $40,000 to $70,0002. The 21st Century Cures Act, enforced beginning 2021, mandates data sharing3, but the practical cost of accessing patient data remains a six-figure undertaking before a single line of clinical software is written. Nebraska’s legislature could change this by funding an EHR sandbox program that lets small, physician-aligned teams access clinical data environments without needing venture capital to walk through the door.
Funding Exists, but the Map Does Not. Nebraska was just awarded $218.5 million from the federal Rural Health Transformation Program4. One of its seven initiatives is a Rural Health Technology Catalyst Fund. But as a medical student building tools, I have no pathway to participate, no point of contact, no process for understanding how these funds flow. The NMA should push for a publicly accessible innovation portal where physicians, students, and builders can see what funding exists and how to apply. Pathways that reward merit, not proximity to the people who control distribution.
Regulatory Fog Around AI in Clinical Practice. No one can clearly answer who bears responsibility when an AI-assisted tool is involved in a bad outcome. A developer cannot obtain malpractice coverage for a product with no defined legal category. Utah is already on the front foot. In 2024, the state created an Office of Artificial Intelligence Policy with a regulatory sandbox, and this year approved Doctronic as the first AI to legally handle routine prescription renewals under state oversight8. Whether or not that is the right model, they are thinking about this early. Nebraska should be too. The NMA is positioned to champion safe harbor legislation protecting physicians, students, and small teams from liability exposure that currently only large corporations can absorb. Audacity to build should not require a liability attorney.
The Barrier No Law Can Fix. We ultimately did not continue building the pre-visit intake tool. The barriers above were a significant part of why. But upon deep reflection, the biggest barrier hasn’t been structural, it has been cultural: medicine’s reluctance to bet on new tools the way Silicon Valley does. I understand why. Physicians are trained in risk aversion, and for good reason: people’s lives are on the end of every decision.
But the American healthcare system is fundamentally broken. $5.3 trillion in 2024, projected to exceed 20% of GDP by 20335. Trust in physicians and hospitals has fallen from 71.5% to 40.1% in four years6. Over 43% incumbent and playing it safe is how we got here and is not how we get out.
Over eighteen months, I reached out to numerous physicians seeking mentorship and co-designers. They told me it was a great idea. Then the conversation stopped. Not because they didn’t care, but because the system gives physicians no incentive to lean into innovation, no protected time for it, no infrastructure to support it.
This year, America turns 250. A quarter millennium built on betting on people the establishment hadn’t credentialed, on trusting that those training to inherit a broken system might be the ones motivated enough to fix it. That spirit built this country. It should not stop at the door of a hospital.
Nebraska talks about cultivating homegrown innovation. It is time to bet on it, on people inside the arena of medicine who build not for market share but because they watched a physician lose half a visit to admin work and thought, “we can fix this.” The legislative fixes above would matter. But none move without physicians who choose to lean in, to co-build, and ultimately treat innovation as part of what it means to care for patients in this moment in history.
*The views expressed above are explicitly my own and do not reflect the official position of the University of Nebraska system.
1 Ptolemay (2024); OmniMD (2026).
2 Thoropass (2025).
3 21st Century Cures Act (Pub. L. 114-255); ONC regulations effective April 5, 2021.
4 CMS Rural Health Transformation Program; Nebraska DHHS RHTP Project Summary, Initiative 7.
5 CMS National Health Expenditure Data (2024); Health Affairs projections.
6 Perlis RH et al., JAMA Network Open (July 2024).
7 AMA Organizational Biopsy National Report (2024).
8 Utah Office of AI Policy, Doctronic Agreement (Jan. 2026).
