Thoughts from Ken Kaufman

So Much to Worry About

2 minute
Healthcare AI

If you are a hospital executive—and if you are reading this, you probably are—then you have no shortage of worries. The worry list is long:

  • Trying to control expenses.
  • Dealing with declining revenue, especially when considered on an after-inflation basis.
  • Struggling with ongoing staffing issues that have no immediate solutions.
  • Solving the longstanding problem of patient access to appointments and service.

And the list could go on and on.

But maybe the biggest concern is one that is not on many worry lists: the remarkable development of artificial intelligence (AI) and how AI is relentlessly pushing into business practice generally and into healthcare more specifically. While the long-term worry is how your hospital will carefully and properly adopt AI inside the business and clinical parts of your organization, the more immediate and short-term worry is whether you, as an executive, understand AI in a way that you can be ultimately useful to your organization.

Full disclosure: I can’t help here much. For me, AI is a pretty big black box. But when I confront this kind of business problem I start reading and learning. One of the most useful AI articles I have come across is “The Optimists: The Full Story of Microsoft’s Relationship with OpenAI,” which was published in the December 11, 2023, issue of The New Yorker magazine. The article was written by Charles Duhigg, a former winner of the Pulitzer Prize.

I am hoping that for your own professional development you will read the Duhigg article, but just in case, here are the highlights:

  • Microsoft has reportedly invested $13 billion in the for-profit arm of OpenAI.
  • Using OpenAI technology, Microsoft has built a series of AI assistants into Word, Outlook, and PowerPoint. These AI assistants are now known as Office Copilots.
  • Knowledgeable commentators say these Microsoft applications are only moderately sophisticated but, honestly, they seem rather remarkable to me. Here are some of Duhiggs’s examples of requests Office Copilot users can make:
    • “Tell me the pros and cons of each plan described on that video call.”
    • “What’s the most profitable product in these twenty spreadsheets?”
  • How about writing projects? Duhigg notes that the Office Copilot can:
    • Create a financial narrative of the past decade based on a company’s last ten executive summaries.
    • Turn a memo into a PowerPoint.
    • Compile a to-do list for Teams video attendees, in multiple languages, after listening in on a meeting.
  • Later in the article Duhigg details the functionality of the Word Copilot:
    • “You can ask it to reduce a five-page document to ten bullet points…[o]r…it can take the ten bullet points and transform them into a five-page document.”
    • It can write a memo based on previous emails you have written.
    • “You can ask, ‘Did I forget to include anything that usually appears in a contract like this?,’ and the Copilot will review your previous contracts.”

Duhigg reports that Microsoft previously acquired a company called GitHub. GitHub is “a website where users shared code and collaborated on software.” Microsoft operates GitHub as an independent division. GitHub has been a very big success and is used by software engineers and, in a short period of time, has grown to over 100 million users.

OpenAI created an artificial intelligence tool that autocompletes software code. Despite reservations at Microsoft, GitHub President Nat Friedman decided to release the GitHub Copilot autocomplete tool. The result has been $100 million in revenue to GitHub in less than a year. 

At the end of the article, Duhigg notes that these early AI business applications are both “impressive and banal.” Banal because they don’t yet live up to the sci-fi predictions for AI and its long-term impact on society.

Honestly, I don’t see it that way. This OpenAI/Microsoft collaboration is only scratching the surface and its potential uses are already endless, waiting to be invented by 100s of millions of users all over the world, including in healthcare. From my seat, the sky is the limit here. Almost anything seems possible.

I hope this summary of Mr. Duhigg’s exceptional article proves useful and advances your awareness of AI’s aggressive and rapid move into day-to-day business—here, through many of the Microsoft productivity programs that every one of us uses every day. In any case, I recommend that you read Duhigg’s entire article. It is most certainly worth your time.

More Thoughts from Ken