AI May Be the Tool to Transform Health Care
Wachter noted that health care was in desperate need of transformation. AI, he said, made a lot of processes better that were already working well. The health care system, he said, was not one of those areas, as most would agree that the health care system was flawed before AI was introduced to the public. Patients have long medical records, and many medical articles are published per day. AI can summarize these and save doctors time, as Wachter noted that doctors have become data-entry clerks in many cases. AI can help address these issues and help doctors do their jobs better. The real game changer in health care will be clinical decision support, where it is integrated into the electronic health record, said Wachter.
He compared AI to Waymo, the driverless taxi that is used in San Francisco. He noted that, just as Waymo made headlines for killing a cat,2 should AI mess up in health care, it could bring disastrous results for the patient and could lead to public mistrust of AI in this field. For the foreseeable future, Wachter sees AI being correct about 70% to 99% of the time, which will require humans to be the safety bulwarks.
“When you think about tasks that humans really stink at, I would put remaining eternally vigilant when you come to trust a technology really high on the list,” he said.
Physicians are best at interpreting AI and need to be the stalwarts for making sure that the patient is not getting the wrong advice for their medical issue.
Why Ophthalmology Is a Special Case for AI
Ophthalmology is a see-through organ, allowing for the ability to train AI on high-resolution imaging of the eye and any eye conditions that could come about. Wachter explained that AI is able to combine the imaging, interpretation, and measurements in ophthalmology, especially as robotic models are catching up to large language models in efficacy. This would allow AI to combine measurements like intraocular pressure and visual acuity and provide a plan for the patient.
AI could also influence the relationships between ophthalmologists and optometrists as AI helps to practice at the top of the eye doctor’s license. Making sure to think through how the algorithm fits into that will be important moving forward. Wachter said AI could likely detect more conditions than the health system can even act on.
Overall, Wachter suggests that health care leaders in the practice should be creating a governance process that works, including ensuring the integrity of the data and creating a system for monitoring. Training staff in the areas of gaps is also important to make sure that there aren’t any potential areas of slipping. Striking a balance between enthusiasm and skepticism is also important, especially as it pertains to managing the anxiety of job replacement.
“I think it’s important to recognize the status quo is not working in health care…In many ways the status quo is immoral. And therefore I think we have to embrace this opportunity to improve our lives but more importantly the lives of our patients,” Wachter concluded.
References
- NBC News Decision Desk poll results. NBC News Decision Desk. September 6, 2026. Accessed October 10, 2026.
https://www.documentcloud.org/documents/28601712-nbc-news-decision-desk-poll-september-ai-topline/ - Knight H. Waymo was thriving in San Francisco. Then one of its driverless cars killed a cat. The New York Times. November 15, 2025. Accessed October 10, 2026.
https://www.nytimes.com/2025/11/15/us/waymo-san-francisco-kit-kat.html
Source: www.ajmc.com
