A just machine?

Donald Fagen was co-founder, lead singer, co-songwriter and keyboardist of the celebrated ‘70s jazz-rock duo Steely Dan. After the band broke up in 1981, Fagen released his debut solo album, The Nightfly, which included the acclaimed track “IGY (What a Beautiful World).” The acronym IGY refers to the International Geophysical Year – a period from mid-1957 to late 1958, when the world’s leading scientists sought to solve humanity’s largest challenges via global scientific collaboration. Fagen’s song exudes the spirit of that era’s shared optimism brought about by technology. The song’s chorus is a brain stain refrain of “What a beautiful world this will be. What a glorious time to be free.” 

Fagen even channeled what he thought people in the 1950s might have viewed as life changing advancements such as solar-powered cities, a transatlantic tunnel, permanent space stations and spandex jackets for everyone (really?). The song’s final verse surfaced the ultimate advancement: 

A just machine to make big decisions,
Programmed by fellows with compassion and vision 
We’ll be clean when their work is done 
We’ll be eternally free, yes and eternally young    

What an idyllic image – outsourcing all of our toughest dilemmas to be solved by a humming metal box with blinking lights. If only things were so simple.  

That song has come to mind lately as I’ve read article after article about the release of OpenAI’s ChatGPT technology. According to the company, ChatGPT can “answer follow-up questions, admit its mistakes, challenge incorrect premises and reject inappropriate requests.” Bill Gates recently called ChatGPT “the most important advancement in technology since the graphical user interface,” and that artificial intelligence (AI) technology “is as fundamental as the creation of the microprocessor, the personal computer, the internet and the mobile phone. It will change the way people work, learn, travel, get health care and communicate with each other. Entire industries will orient around it. Businesses will distinguish themselves by how well they use it.” 

Impressive claims as bold as undersea train tunnels and spandex jackets. Color this luddite skeptical.

Now I readily admit AI has immense potential, especially when combined with ambient speech technology, but the hyperbolic vision painted by Microsoft’s founder seems a bit fanciful. Not so long ago our friends at IBM proclaimed their humming metal box Watson, “capable of reading 800 million pages per second,” could consume, contextualize, learn from and synthesize the entire world’s medical literature, documentation and artifacts. IBM’s Global Brand Manager Ann Rubin, even assured us, “You can outthink cancer, outthink risk, outthink doubt, outthink competitors if you embrace this idea of cognitive computing." 

Watson displayed great promise by appearing on the popular gameshow Jeopardy and handily defeating former champions Ken Jennings and Brad Rutter, but by early 2022, IBM had completely abandoned the project and sold Watson off for parts. Such promise.

While many people are fine to use self-checkout stations at the supermarket, the vast majority still prefer to have humans conduct certain high stakes activities. Imagine yourself travelling home from a business trip with a two-hour flight to your destination. Would you choose a sleek, shiny Tesla aircraft with an empty cockpit or an older 747 driven by a former Naval aviator with 30 years’ experience, who is looking forward to dinner with his family as well? Most people would opt for the pilot with a pulse. Why? The answer relates to what author Nassim Taleb describes as, “skin in the game.” In our air travel example, the pilotless plane will likely get us home safely, but one computer glitch or incorrect algorithm calculation could prove disastrous. Make no mistake, our human pilot isn’t infallible either. He has slower reflexes and poorer vision than in younger years. He may have high blood pressure and perhaps didn’t sleep well the previous night. But here’s the difference – our human pilot has just as much to lose as his passengers. He has skin in the game.

The same concept applies to the use of technologies like ChatGPT in health care. Reports estimate that 10-15 percent of all medical diagnoses in the world may be incomplete or incorrect. Not a great batting average. However, the promise AI brings to health care and the reality it brings are still miles apart … at least for now. Perhaps one day “a just machine” will exist to make medical and other weighty, life decisions for us. Until then, the best we can expect is a correct answer to the “Jeopardy” Daily Double.

Clark Cameron is manager of payer market strategy and development for 3M Health Information Systems.