Stay Updated

AI Won’t Save Healthcare If Pointed in the Wrong Direction

Ben Bobrow, MD, FACEP, is a Professor of Emergency Medicine and VP of Healthcare Innovation at the McGovern Medical School at UTHealth Houston. Dr. Bobrow is the inaugural Chairman of DocGo’s Medical Advisory Board.

Artificial intelligence has quickly captured the imagination of the healthcare world. The conversations are constant, the headlines are bold, and the expectations are climbing. I recently attended an exciting and motivational conference hosted by the University of Texas health system. It brought together hundreds of top data scientists, clinical and administrative healthcare leaders to explore AI’s role in transforming medicine. Presenters showed how AI can revolutionize protein design leading to the rapid discovery of pharmacological agents to cure various important diseases like dementia and different kinds of cancer. What I saw captured my imagination, and the seemingly limitless potential offered by AI energized me.

Beneath all that excitement, however, I recognized something familiar. We were all exhibiting recency bias. The newest technology, the most current research, the latest breakthroughs (and failures) tend to dominate our thinking. It is easy to get swept up in the breathtaking potential. I do it too. This is why it is so important to regularly return to first principles. Rather than focusing on chasing distant breakthroughs, the more attainable promise of AI, I believe, lies in addressing the most persistent problems right in front of us using tools we have in hand.

The next great advancement in healthcare might not be a new miracle drug, but rather our dramatically improved ability to deliver existing therapies more equitably, more efficiently, and to more people. There are so many unacceptable disparities in our system where this thinking applies. For example, Texas is home to over 4 million people living in rural areas, spread across a state that spans more than 268,000 square miles. That’s larger than the population of over 18 U.S. states. With so much ground to cover and many patients suffering due to their limited access to care in many regions, how could AI help? One way AI can transform rural healthcare is by expanding access through virtual care, remote monitoring, and intelligent triage – helping patients receive timely, high-quality care without traveling long distances. AI can also support overburdened providers with diagnostic assistance, predictive analytics, and administrative automation – ultimately improving outcomes, reducing costs, and strengthening rural health systems. 

Yet the current trajectory of AI in healthcare seems headed in a different direction. We are already seeing insurance companies deploy AI to process claim denials faster, while patients turn to ChatGPT to write more sophisticated appeals letters. The technology makes both sides more efficient at fighting each other, which contributes to a more adversarial system where the people in the middle suffer. It also transforms innovation into a tool that makes old frustrations more efficient rather than solving underlying problems. We need to use AI to reduce conflict, not multiply it.

I keep returning to the idea that physicians should be able to practice medicine without being overwhelmed by administrative tasks. I have seen firsthand how early electronic medical records (EMRs) disrupted that goal. EMRs promised to streamline documentation and improve care coordination, but they often introduced extra layers of friction and took clinician time away from patients. Many of those systems were built without a clear understanding of clinical realities, turning doctors into data entry clerks. The frustration was palpable: “Just one more click,” “two more clicks,” “three easy clicks.” That crushed physicians and nurses alike with tiresome inefficiencies. We’re seeing echoes of this with AI now, and some providers remain wary of adoption, wondering whether it will actually improve things or make them worse. Because AI carries even more power than those early digital tools, it demands even more care and direction in its implementation.

What I hope to offer is not a novel idea but a reminder. AI can help physicians return to the work that drew them to medicine in the first place. It can reduce disparities, strengthen the care relationship, and expand access for the millions of Americans who live in regions with too few providers. These are not abstract benefits – they are measurable and immediate.

Every discussion about AI in healthcare should include a simple question: Will this help more people receive the care they need, when they need it? If the answer is no, the technology may be interesting, but it is not what we need most. 

If the answer is “yes,” then AI has the potential to become a transformative tool, helping practitioners put patients’ needs first. The opportunity and the challenge ahead is to focus on meaningful outcomes over technological razzle-dazzle. It won’t be easy, but the rewards have the potential to be literally life-changing.