How will machine learning impact you?
I saw an interesting headline in the news this week, which read:
“Microsoft’s cloud boss says the company doesn’t want to compete with doctors.”
This statement is coming as Microsoft and other Cloud providers are making their way into the healthcare space. Microsoft recently announced the acquisition of a speech recognition company that doctors use to keep notes on meetings with patients. The move will provide Microsoft an expanded Cloud offering and, perhaps incidentally, access to lots and lots of data when it comes to healthcare and diagnosis.
Artificial Intelligence in healthcare
Like any knowledge-based industry, doctors should be wary of the long-term impacts of artificial intelligence on their industry. For the most part they are, which is exactly why Microsoft is putting out this messaging. “We are here to help!” Radiology, the medical practice of diagnosis through highly trained analysis of visual scans, is probably on the very frontline, in terms of disruption by AI. From 2015 to 2020, adoption of AI in radiology clinics has gone from 0% to 30%, and it is very likely this will increase significantly. The economic upside of being able to increase the capacity of highly trained workers, and then potentially replace them, is simply too large to ignore.
Where’s my self-driving car?
Those of us who have been following Artificial Intelligence long enough have been sounding the alarm for years about human drivers being replaced by machines. Back in 2018, I wrote about the gains in time people would be able to unlock, when they no longer had to pay attention during their commute. But now, over three years later, where’s the Driver Apocalypse?
- There are still 1 million Uber drivers in the United States and 3 to 4 million world wide.
- There has been a shortage of long-haul trucking drivers since the 1980s, and that gap still exists in 2021.
With the potential financial gains to be made replacing drivers, I (and many other futurists) imagined whole fleets of self-driving cars and trucks on the road right now. And I still do believe that will happen…eventually. But, change is slow until it’s not. For those well-heeled enough to afford a Tesla, this demo of Autopilot is pretty darn impressive:
Two models of displacement
While the speed of the rollout of artificial intelligence may be slower than predicted, change is coming to every industry. It will not be years, but over decades, humans in every industry will first be augmented and then replaced. It’s happened before.
The first domestic dishwasher with an electric motor was invented and manufactured by Miele back in 1929, but it was only in the postwar boom of the 1950s that they were successfully marketed and sold to people. By the 1970s, dishwashers became commonplace, but even today, only 75% of American households have a dishwasher. If this is our model, we still have a ways to go. It’s likely that we might not see 100% adoption even a century from now. I believe this may be the case for consumer adoption of driverless cars. Since it is a matter of convenience, some folks will always choose a lower cost option.
However, innovation isn’t always this slow. When Uber came out, for example, it massively disrupted the New York City taxi industry which has a system of medallions authorizing people to drive cabs in the city. These medallions were seen as strong assets. They grew in value from $25,000 in 1962 to $600,000 in 2010, peaking in 2013 at a value over $1 million. Uber came to New York City in 2011, and between 2014 and 2015, the value of taxi medallions had dropped by 50%. In 2019, 16 medallions were auctioned off, with three selling for around $130,000 and 13 finding no buyers at all.
Dishwasher or taxi?
Today, everyone, doctors included, should be asking themselves two questions:
- Can a machine do my job?
- Will a machine cost less?
The answer to the first question is probably yes, especially for knowledge workers (doctors, lawyers, programmers, etc.) The answer to the second question may be no, especially in the case of human-facing jobs (retail, service sector, etc.) or jobs that require complex labor across multiple locations (making beds in a hotel, cleaning rooms, etc). One barrier to replacing cab/Uber drivers, for example, is the simple act of helping people with their bags.
For most of us, there is at least some work that we do which a robot could do and do for less. In the case of Microsoft, you can bet they will be identifying and monetizing all of the things they can within the healthcare space. At first, there may be a lot of time savings they can unlock for doctors. But read that headline carefully: “Microsoft’s cloud boss says the company doesn’t want to compete with doctors.” The announcement doesn’t say they can’t compete with doctors, or even that they won’t. For now, they are happy to augment and support. But make no mistake: the machines are coming.