The other day I was walking by a construction site. I wondered what they were building. The fence around the site was covered by blocking material, and there was a sign with the construction company’s name on it. I walked up to the sign and saw a picture of a gleaming building, with a central plaza filled with smiling, busy people. It looked pretty cool. But then I peeked through the fence and saw…a giant hole. A giant hole in the earth, with mounds of dirt around it and scattered piles of building material. A few folks in fluorescent vests and hardhats were walking around.
I imagined the conference rooms where this conceptual rendering was shown, gathering interest and excitement for the project. Then I thought of that wide gap between those rooms where the vision was created and the gritty hole in front of me, where nails are going into wood and concrete is being poured.
This image came to mind recently when listening to an excited podcast host talking about the potential for AI in healthcare. Less burnout! Better quality! Lower costs! We’re going to fix the healthcare workforce shortage! It is exciting, (it’s the area I’ve chosen to work in, after all) but boy, it’s going to be a lot of work. And we’re going to have to pay attention to so many small details.
The vision is necessary for creating goals, gathering support, and igniting inspiration, and that’s the realm of the dreamers—the people in suits in conference rooms—creating plans and setting things in motion. But they aren’t the ones who determine how things look in the end. That’s decided by the builders—those people in fluorescent vests, asking the hard questions, creating the safeguards, getting the details right. The final result depends on adding up all those little actions and decisions, and the end might or might not resemble the conceptual rendering vision. It might fall short, or it might end up somewhere else entirely. After all, when the internet first reached the public eye, who guessed that in a few years we’d be complaining about 30-second videos shifting national elections?
Personally, I find myself to be about 20% dreamer, 80% builder. I can agree with the optimistic vision, but I very quickly turn to questions of how this is actually going to work. Here are some of the visionary assertions I’ve heard recently, and some questions that builders might want to ask:
Every doctor should be using AI with every patient visit
The scribing software—taking the conversation and turning it into a medical note—is the most promising technology that might fall under this statement, but we still don’t quite know how to ensure the note accurately reflects the conversation, or what the effect will be of having a doctor act as a proofreader rather than a writer.
We just have to remember to “keep the doctor in the loop.”
Similar to the scribing situation, that sounds good, but it depends on all sorts of design decisions around how we put the doctor “in the loop.” How do we know when we should trust the software, and when we need a doctor to go another way? And how do we make sure doctors are really empowered and enabled to steer the care, rather than just being a rubber stamp and a bearer of liability?
We can have patients talk to a chatbot first.
It’s pretty jarring to be talking to a bot and then only later realize that you aren’t talking to a person. How do we preserve the human-to-human relationship, when that’s what is needed? How do we monitor these chats for quality and accuracy?
AI can take away the drudgery and do a lot of paperwork for you.
Except that sometimes, the paperwork is there for a reason, commonly to contain costs. It’s a barrier to spending money. If we make that less of a barrier, those who wish to control spending will just add others—probably using AI—until we’ve got an AI battle on our hands. More than that, sometimes the paperwork is a sign of inefficient processes. If we automate these using AI, we aren’t fixing those processes, we’re just taking the most advanced technology we have and using it to retroactively neaten a process that probably could’ve been made less messy from the start.
AI can start in areas with poorer quality healthcare. It doesn’t have to be the best care, it just has to be better than what they have now.
This may be a good start and it might be the first bar, but we don’t want to stop there. We have the opportunity to create better care for everyone, not create a tiered system, where those who start with fewer resources are subjected to incomplete or untested tools, because it’s better than what they have. It’s going to take some careful work to balance helping people today while refraining from solidifying an ongoing lower standard for segments of society or the world.
We see the same dynamic between the vision and the work in other parts of life. Take parenting. The dream is to turn a child into a happy, independent adult. But the parents are often the dreamers, and the dream only comes true from the work of the builders—the kids themselves—who have to do the homework, attend the practices, and write the applications to bring that dream to life. And just like with AI in healthcare, it’s possible to set unrealistic goals or demand work that actually undermines the dream we’re shooting for.
I know that focusing on the gritty reality and the problems without obvious solutions is not “sexy.” It doesn’t make sales, doesn’t bring in millions of dollars in VC investment. But just living in the world of visions and dreams doesn’t get the job done. We need both dreamers and builders on the team, or we need people who can play both roles. Consider whether you are more of a dreamer or builder, and if you could use a little bit more of the other side. How do we bring both perspectives together—not just to imagine a better future, but to actually build it?
If you know someone who might like this post, please feel free to share it with them. If you’d like to receive these posts (and a few bonus posts) by email, subscribe here.
Red Goggles 