For future AI, the prompt is you

Here’s how to navigate the future of healthcare with AI by using strategic principles that align the workforce and more.

This article is part of AI BEYOND the Hype - March/April 2024 COVERstory.

In their book, “Exponential Organizations 2.0,” Peter Diamandis and Salim Ismail talk about the fact that the next billion-dollar start-up will be created by three people – a visionary CEO, a product leader and an operations leader. The rest, they surmise, is going to be done by AI.

AI is an exponential technology, following the same characteristics of Moore’s Law – doubling in performance and sophistication every 18 to 24 months or so. And that’s why, every time we see a new release of the next language model, it has wondrous new capabilities and applications.

Now consider the application of AI in healthcare. If there were ever an opposite of Moore’s Law, healthcare would be a good candidate. It accounts for spending of $4.3 trillion annually and rising – accounting for nearly 20 percent of the GDP in the U.S. Yet quality is declining, access to care remains spotty, and the number of uninsured and under-insured keeps growing. Something must change –– and that change starts with you and your imagination.

Even Einstein said that imagination is more important than knowledge, and it’s rumored that the theory of relativity started with a thought experiment. That is key to what we need to consider –– whether you’re a technology leader that’s passionate about AI, a P&L leader that believes in the business model implications of AI and wants to explore what’s next or the CEO who simply wants to take the enterprise to the next level.

Here are some design principles I’ve learned over the past 20-plus years.

Of course, some of these principles can be considered table stakes. For example, every AI initiative must begin with data, data quality and data integrity. And, of course, data interoperability and model interoperability are paramount.

AI literacy for all

We must build in ethics, eliminate bias and ensure transparency. That’s a given. But we also must upskill the workforce – not only the technical workforce — in AI literacy. And there must be alignment –– top-down, board-level and bottom-up alignment. AI initiatives must be treated like any other change program –– with cultural aspects front and center.

Some of these principles are less obvious, such as “purpose driven.” As an example, consider a company whose purpose is to simplify the business of healthcare. Now, let’s distinguish that purpose from the company’s goal, which is to maximize value and ROI for its clients. As new initiatives are taken on, the organization must ask itself how that AI initiative is going to help it achieve its purpose. This exercise activates the entire enterprise –– making AI everyone’s business, instead of just making AI the business of a select few data scientists and engineers.

Portfolio-based approach, not moonshots

Scaling AI for healthcare requires a measured, portfolio-based approach. Sure, we all love “moonshots” and ambitious projects, but we can’t start there. Singles, doubles and a more pragmatic portfolio of initiatives is the way to go.

Take a simple example: If, years ago, Tesla had said they wouldn’t ship a single car until autonomous driving was perfected, they would likely be out of business now.

When you evaluate the typical portfolio of AI initiatives today, it’s stacked in favor of the moonshots. AI to cure cancer. AI to replace humans. Ambitious? Yes. But a more realistic approach often yields better results.

AI operating model

Another key design principle is the operating model for AI. We cannot simply take AI and drop it into a company’s existing organizational structure.

There’s a solid precedent for this – nearly 200 years ago, as we moved from steam power to electricity, electric motors enabled us to change the factory floor and organize the factories around the workforce. This eventually led to the production line, generated considerable wealth for the nations that took part in that revolution, drove down consumer prices and created the modern middle class. So a similar operating model mindset for AI will be paramount to getting it all right.

Platforms, not point solutions

Lastly, we need to think in terms of platforms rather than point solutions. For example, consider exiting a conference and hailing a cab. Perhaps the driver is using Waze, which uses AI, and gets you to your destination quickly. This is a solid point application of AI.

Now consider a ride share app. The ride share platform, in the background, weighs the mass supply-demand optimization given the conditions and time. It is likely predicting that the 15,000 people at the conference will be exiting the venue around 4 p.m., and therefore configures supply to be available so that no one must wait longer than two or three minutes for a ride.

It considers optimization, supply chain, platform, consumer experience, driver experience, driver payments –– all of which is optimized by AI, vs. the point solution of taking you from point A to point B.

Healthcare is no different. Consumers get care. They go doctors and hospitals, they get labs done, they get therapeutics and drugs. But all those interactions happen in silos. None are interoperable.

But what if we created a platform where we, as consumers of care, were more ubiquitously connected and integrated with all the supplies of care? The data from every interaction between a care recipient can be tracked, and from that data insights generated. And with those insights we can start to make healthcare much more predictive, proactive and personalized.

Biology as the new asset class

Now, what if we could apply all these same design principles to ourselves? What if you, as the asset manager of your own biology – your medical and claims data, benefits data, purpose, social factors, all of those things – could prompt a large language model and teach that model everything about you with a singular purpose – to improve your health and well-being?

Whether you’re the technology leader, the P&L exec, the CEO or simply someone who wants to use AI to improve your health, in all those cases, the prompt is you.

Rajeev Ronanki is CEO of Lyric.

Return to AI BEYOND the Hype - March/April 2024 COVERstory.

More for you

Loading data for hdm_tax_topic #healthcare-equity...