Which Jobs Will Machines Take Over? Movie Critics, Doctors, Truckers... | Big Think

So machine intelligence includes two main areas. One is high-performance computing; all the chips you need to do the complex math required, and then artificial intelligence. Within artificial intelligence, all the breakthroughs are happening in the area of machine learning, and machine learning includes the ability for computers to think, learn, and act on their own.

There are a lot of great examples throughout business, government, and the nonprofit world as well. For example, in the U.S. government, machine learning is completely changing how census workers visit houses to collect the census data. These enumerators in the past would just use their own judgment on what routes to take and when to visit houses, and, of course, they often found people not home.

In the upcoming 2020 decennial census, machine learning will actually give them the best routes and the best predicted times to find when people are going to be home. When people ask me if machine intelligence will affect them, I turn the question back around and say, “Could you imagine your life or your job without the Internet?” And most people will say no. And I say, that's how machine intelligence will become for us.

It's already involved in our personal lives whether it's Amazon recommendation engines or Netflix recommendations, but it will continue to spread not only in our personal and entertainment lives but also in our day-to-day work lives as well. One example is IT departments in corporations. Many companies spend the majority of their IT budget on IT operations, operations and maintenance, and it's a huge budget line item.

Machine intelligence is now able to perform most or all of those mundane routine activities. That frees up an entire skilled workforce to focus on creativity and innovation beyond the jobs that they're doing today. So, machine intelligence is not about completely removing the person from the equation; it's about machines and humans working together.

We never recommend that you completely abandon your judgment as a leader. Many times machine learning is wrong, and as leaders we need to recognize that and know when to ask questions and how to adapt. So judgment is still absolutely critical in this equation.

An interesting example is automated trucking company Otto. It was formed by Google engineers and within ten months was bought by Uber for $680 million. Now what would make a company worth $680 million in ten months? Well, Otto developed a cab-top device that sells for $30,000, and can fully automate any 18-wheeler truck built after 2013.

In fact, they've already made their first delivery using a self-driving truck: it was 50,000 cans of Budweiser beer, last year. Right now the truck can't handle every situation; it's really great on the highways, but it has problems in bad weather and on city streets, so for now the drivers continue to ride in the cab, but when they're in on the highway, the driver is free to conduct whatever activities they choose—and I love the story of a driver doing yoga while going down the highway—and then the driver takes over in difficult situations.

There is no question that artificial intelligence will replace certain types of tasks. There are tasks today that machines are better at than people; tasks like organizing, remembering, finding patterns. So those kind of tasks will be replaced by machines. But people are still so much better at creativity and framing problems and reasoning.

So what will happen is that machines will replace some of the rote tasks that people are doing today, and it's up to us then to figure out as leaders: what do we do with that talent, and how do we take advantage of the opportunity that that gives us? So it's not a matter of machines putting people out of work across the board; what goes hand-in-hand with that are the new businesses that it will create, the new industries that it might spawn, and a whole new set of demands that don't exist today that people will be able to go into for a second or third career.

One example of this is the Data Science Bowl. Every year we put on the Data Science Bowl with Kaggle, and we choose a really hard social-good problem. Kaggle puts it out to its community of a million data scientists around the world, and they work as volunteers to solve the problem. Last year the challenge was around heart health, and this year it was around new ways of predicting lung cancer.

And of the top ten winning teams, many of them had no background in the medical field whatsoever prior to this competition. They were able to learn the medical knowledge they needed through tutorials, and because they came at it from different angles and different perspectives, they were able to make breakthroughs in the lung cancer challenge. The top ten algorithms will improve by ten percent the accuracy of predicting lung cancer.

In fact, one of the winning teams has already secured venture capital funding to turn that algorithm into a new business. So that's an example of a business that didn't exist previously and is launched because of machine intelligence. As a society, we need to make sure we help folks whose jobs will be taken over by AI because there will certainly be tasks that no longer require a person to do them.

The answer to that dilemma is to retrain those folks into different skill sets that can be applied to the new industries and the new companies and the new jobs that are created. So the best policy position that we can take as a nation is to invest in retraining those workers. It doesn’t help to deny the problem or to think that it’s “way out” when it’s already happening today.

We need to be proactive about retraining that workforce so that we don’t have a class of people left behind. And by the way, this doesn’t apply only to blue-collar type jobs. For example, doctor—their jobs will be completely disrupted and changed. We will no longer rely on a doctor’s knowledge or experience to diagnose and treat diseases.

Instead, doctors will become interpreters of models. The legal profession will be disrupted. No longer will we require people to do research on precedence and case law; machines will be able to do that. So it affects every walk of life and every industry, but we need to invest in the people who are displaced so that we take full advantage of that talent to use toward greater ends.

One thing I'd say about barriers is that anyone can learn machine intelligence. Many people feel like this is something they had to major in in school or they have to work at a tech company to become an expert. That's no longer true; many of the leaders that we interviewed for our book had no background in machine intelligence whatsoever before they took on these projects, and so anyone can and should learn about machine intelligence: it will be that much of a game changer.