I expect a lot of people to be bothered by ChatGPT Health. Of course you shouldn’t use AI as a replacement for a human doctor. And what about privacy?
AI for health questions has actually gotten really good in the last year. It’s better at explaining lab results. It’s better at understanding medications. It’s just generally better at pointing us in the right direction about complex topics.
When my mom got sick last year, I spent many days in the hospital, in the ER, at rehab facilities, talking to nurses and doctors. It’s what inspired me to make Micro.blog free for nurses. Doctors and nurses are trying their best and sometimes they do work miracles, but they are overwhelmed by the system. Too many things fall through the cracks.
There are a couple of significant problems that could be addressed by AI:
- Health providers do not have time to explain to us everything they have learned from years of schooling and hands-on experience. We know next to nothing about medicine, so we don’t know what questions to ask or when to help a doctor with important context.
- Health records are a disaster. They are fragmented across multiple providers. They are wildly incomplete. Even when a patient’s records do contain relevant details, it is too much information for a human to quickly make sense of. And with phones and watches, we’re collecting more health data than ever, but it goes unused.
I can’t count how many times I had to explain my mom’s medical history to a new nurse or doctor, just to get them up to speed. They need the full picture, not just a quick glance at the latest vitals and blood work. And at the same time, doctors have to spend precious minutes summarizing a procedure or medication in terms that we can understand.
I don’t have even the slightest worry that AI will replace doctors. There will always be more work than they can handle. Technology should be an amplifier, letting professionals do their jobs more effectively, so they can focus more time on the things that only they can do.
We are only at the very beginning with AI for health, and it can be scary to move too quickly. But in a decade, if this is properly woven into the fabric of health care, it will save lives. I’m not talking about breakthroughs and cures, at least not yet. There is mundane work that breaks down under the frenzy and scale of modern health care — context, communication, and surfacing the right details at the right time — problems that LLMs are well suited to fix.
