By Professor Aymeric Lim, M.D.
0
Share
Newsweek is a Trust Project member
See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.
There is a familiar story being told about artificial intelligence in health care. In this story, an algorithm reads a scan more accurately than a radiologist, a chatbot triages a patient more efficiently than a nurse and the hospital of the future is a quietly humming data center with a few humans wandering its corridors. It is a tidy narrative. It is also, in my experience, almost entirely wrong.
After years of building, deploying and sometimes retiring AI tools at the National University Hospital, Singapore (NUH), I have come to believe the most important question is not what the algorithm can do, but what the algorithm can allow people to do that they could not do before. Technology is the easy part. The harder, more interesting and ultimately more transformative question is one of leadership: How do we build a workforce that uses AI as a collaborator, not an oracle?
Our answer begins with a number: ten percent. Our ambition is to train ten percent of staff in every job group — not just doctors, not just IT — in the skills of innovation and AI literacy, including cleaners, pharmacists, schedulers, physiotherapists and finance officers. Ten percent in each group, because that is the threshold at which a culture begins to shift. Below it, innovation is something “those people up there” do. Above it, innovation becomes a verb that everyone uses. We are not trying to produce ten thousand AI engineers. We are trying to produce ten thousand people who can recognize a problem worth solving and have the confidence to bring it to a team that can solve it.
Read More on Community Forum
To support that ambition, we have established the Kent Ridge Office of Innovation, or KROI, which works through three deliberately complementary pillars. ASCEND focuses on continuous learning and skill development to keep up with a technology that does not sit still. NEXUS focuses on collaboration and process optimization because most AI failures in health care are not technical failures; they are workflow failures. And FORGE focuses on technological development, adoption and integration, the difficult, unglamorous work of turning a promising prototype into a tool that survives contact with a real ward at 2 a.m. We are proud to share that to date, 47 percent of our more than 9,000 staff have undergone training in combining AI with clinical and operational improvement.
Alongside KROI, we have launched a dedicated Innovation Hub: a physical and intellectual space where staff can test ideas, access emerging technologies and collaborate across disciplines. The Hub matters because it makes innovation tangible. You can walk into it. You can break things in it. You can leave with a working prototype and a list of three colleagues from other departments who want to help. The most underrated infrastructure in any digitally ambitious hospital is not the data center; it is the room where a nurse and a software engineer can sit across a table from each other and understand what the other one is saying.
This is where the parallel with mission command becomes important. The same philosophy that guides our clinical leadership guides our approach to AI: clear intent, decentralized execution. We do not believe a small central team can identify, prioritize and deliver the hundreds of meaningful AI use cases hidden inside the daily routines of a more than 9,000-person organization. We believe our frontline teams can if we give them the framework, the tools and the permission to try.
...
The result is a digital workforce that is genuinely a workforce, not a procurement. We are building a portfolio of AI agents to handle the mundane work, such as drafting routine documentation, summarizing long case notes, surfacing the next-best action from a sea of guidelines, reconciling rosters and chasing missing data. None of these tools replaces a health care professional. All of them give a health care professional back the one resource they cannot manufacture: time. Time to think. Time to listen. Time to be present with the patient in front of them.
There is a counterintuitive insight at the heart of this. Successful AI adoption requires more human leadership, not less. Algorithms cannot decide what matters. They cannot weigh a competing set of values. They cannot stand at the bedside of a frightened family and explain a difficult choice. Every meaningful use of AI in health care ends with a human being making a judgment, and that human being needs to be supported, trained and trusted. Otherwise, we will simply have automated our worst habits at scale.
This also changes what we look for in our leaders. It is no longer enough to be a skilled clinician or a competent administrator. The leaders we are developing need to understand what AI can and cannot do, how to interrogate a model’s recommendation, when to override it, when to escalate and when to admit that the algorithm has spotted something the team missed. That is a new literacy, and like all literacies, it has to be taught, practiced and renewed.
The institutions that will get this right are not the ones with the biggest AI budgets. They are the ones whose people feel like co-authors of technology, not passengers on it. That is a leadership project, not a procurement project. It is slower, messier and far more human than the glossy keynote version of digital transformation. It is also the only version I have seen actually work.
If we do this well, the hospital of the future will not look like a data center with a few humans in it. It will look like a hospital staffed by people who, thanks to the quiet work of a thousand algorithms in the background, finally have enough room to be the health care professionals they trained to be.
Aymeric Lim is a member of the Newsweek CEO Circle, an invite-only executive community of subscribers.
View 0 Comments
Submit Correction
View Editorial & AI Guidelines