Artificial intelligence and automation technologies are bringing massive change to labor, the economy, ethics, and society. The question now is how to harness those technologies rather than being overwhelmed by them—or overwhelmed by those who embrace them sooner. We break down the 20-year outlook with Laura Schultz, director of fiscal analysis and chief economist at the Rockefeller Institute of Government.


Laura Schultz, director of fiscal analysis and chief economist, Rockefeller Institute of Government

Learn More:

The Impact of Artificial Intelligence on the Labor Force in New York State


  • Transcript

    Transcript was generated using AI software and may contain errors. 

    David Bowman (2001: A Space Odyssey) 00:00

    Open the pod bay doors, HAL.

    HAL (2001: A Space Odyssey00:04

    I’m sorry, Dave, I’m afraid I can’t do that.

    Eldon Tyrell (Blade Runner) 00:07

    Commerce is our goal here at Tyrell. More human than human is our motto.

    Unknown 00:11

    It’s not a human mind. I mean, look at this, like neurons firing. Artificial intelligence.

    Kyle Adams 00:39

    Welcome to Policy Outsider from the Rockefeller Institute of Government, where we explore and explain how public policy shapes our everyday lives. I’m Kyle Adams, communications director at the Rockefeller Institute. Today, we’re talking about robots, more accurately about artificial intelligence and the effects it will likely have on labor, the economy, ethics, and society. The Rockefeller Institute and SUNY recently hosted an AI policy forum, where we gathered experts in education, industry, and academia, to talk about how we can best transition to the impending age of AI. To help us make sense of it all, we’ve invited Rockefeller Institute senior economist and director of fiscal analysis, Laura Schultz, to discuss her takeaways and the future of AI in New York. Laura, welcome, and thanks for coming.

    Laura Schultz 01:29

    Hello, happy to be here.

    Kyle Adams 01:31

    Before we dive into the forum that we hosted, can you give us a brief overview of artificial intelligence, starting with what it is?

    Laura Schultz 01:41

    Sure, maybe it’s good to start with human intelligence. Two of the things that we think of humans as being able to do is perceive things really well, which means that if I were to sign my signature a thousand times, each one of those would be different. But a person would be able to look at it and go, it’s all the same, whereas the computer wouldn’t really be able to pick up on that. Then the other part of human intelligence is being able to process a lot of information at the same time and come to a conclusion really quickly. Computers haven’t been able to do that quite yet either. Artificial intelligence is a set of technologies that are able to do things that we thought only humans were capable of until just recently, for instance, perception.

    Kyle Adams 02:28

    One of the more immediate or common uses that comes to mind for me is self-driving cars, they will need to be able to recognize pedestrians and do visual calculations that we really think only humans can do and make those judgment calls.

    Laura Schultz 02:44

    Exactly. I think the judgment call is part of it, too. An autonomous vehicle can look and see another car coming and can see whether something’s red to green. But being able to detect all of that and then also figure out how to act appropriately is another big part of artificial intelligence.

    Kyle Adams 03:03

    This is going to probably touch a lot of our lives in ways we can’t quite predict yet. You’ve recently written for the Rockefeller Institute about how artificial intelligence and automation will likely affect the workforce in New York in the next 20 years, and you’ve come up with 53 percent of the workforce is likely to be affected by this, jobs lost or changed due to AI. Now, I find that description lost or changed to be pretty important, and possibly offering a very wide range of effects. So from somebody being replaced on the factory floor by a robot to somebody just having to learn a new software at their office, is there any way to narrow that range when you say lost or changed for 53 percent? Where’s the cut-off and the effect that somebody could see in their job?

    Laura Schultz 03:52

    I think it really highly depends on the job. We can think about, for instance, a bank teller. We now all have access to mobile aps that we can use from our phones or for computers that allow us to take a picture of a check and cash it from the convenience of our home. That was a job before we had to have a bank teller do, so the nature of workers in service industries like banking, at that level could be eliminated or seriously changed. Those people may now be processing loan applications and talking to people about more complex interactions. Other jobs like attorneys or doctors are going to see the nature of their work really change. There are going to be tools that allow them to do their work a lot more efficiently. A doctor instead of spending 15 hours of their week doing paperwork, software and programs and devices are going to be able to process all of the notes they took during meetings and automatically generate all the paperwork they already need. That means that those professions are going to see the nature of the work change. They’re going to be able to focus in more on what is their specialty and less on the minutiae of what they’re doing.

    Kyle Adams 05:11

    We’ll return to the implications of all of that. But for now, I’d like to return to your initial blog, which will be the first in a series as we examine and track the implications of artificial intelligence in New York. How did you make those calculations and projections for the next 20 years?

    Laura Schultz 05:30

    There are experts out there who know the science and know the engineering and know the technology behind artificial intelligence. They worked with a group of economists to look at 800 different professions and the tasks associated with those professions to calculate the likelihood that a computer could do some of those tasks. They are the ones who assigned the likelihood calculation. What we did was we took those probabilities and compared them to the New York State workforce to find out what could happen for us.

    Kyle Adams 06:04

    What were the three most effected fields that you found?

    Laura Schultz 06:08

    The ones in New York that are going to be most impacted are office and administrative support. Right now, we have about a million and a half people working in New York in that job. That’s why it’s going to be highly impacted. Sales jobs are going to possibly be automated. We’ve already seen that as we go into stores and we do automated checkout places. Then finally, food preparation and serving. Restaurants are likely to see high levels of automation in the near future.

    Kyle Adams 06:39

    I’ll jump right into the big question here, what happens to those workers?

    Laura Schultz 06:42

    Well, what’s interesting is those are a lot of jobs that firms are having a hard time filling. Restaurants are having a hard time finding people who want to work as servers or who want to work in the kitchen. Stores are having a hard time finding retail associates. I think that part of it is that automation is filling the gaps that the workers aren’t there for. I think that’s where we’re going to see AI first.

    Kyle Adams 07:12

    Two weeks ago, at the Rockefeller Institute, we hosted a forum with State University of New York on AI policy. We had two excellent panels filled with experts from industry, government, and education, discussing how we can make this transition to a world full of AI basically. How to not be afraid of that, but to harness it. I would like to hear your major takeaways from that forum. Let’s start with a piece you wrote for us recently that AI will be big for New York. Why do you say this?

    Laura Schultz 07:45

    Well, New York is a hub for information and data and major tech companies. All of those companies are leading the AI space. We have IBM, which is a major chip computer server manufacturer. They’re going to be the ones developing the technology behind AI. Then I think New York is an important place for AI to happen, mainly because we have the major industries and the major companies leading AI right here. One of the top companies in AI is Amazon. They’ve just announced that they’re building a second headquarters in New York. We have IBM, which is a major leader in AI, both the hardware and the software and the data analytics fields, here as a corporate leader in AI. Google just announced that they’re expanding their headquarters in New York as well. All the people who do AI are in New York, so we need to take advantage of that.

    Kyle Adams 08:51

    A point that I found interesting from the panel, and maybe it’s because I didn’t know it before, it was just that one of the biggest fields for AI or biggest applications for AI right now is in financial trading. Again, I tend to think of the things that you see in the news more like self-driving cars. I assume that’s just doing faster transactions, smarter transactions, taking the human traders out of the equation?

    Laura Schultz 09:14

    I think it’s in hopes of predicting what’s going to happen. If you can see what’s going on with mortgages, maybe you can understand what’s going to go on in the mortgage crisis. If you can use more information about consumers, you can better predict what’s going to happen to consumer markets. Oh, I think that they’re trying anything and everything and trying to really understand what data can provide them.

    Kyle Adams 09:37

    Seeing trends and massive amounts of data that people probably can’t process.

    Laura Schultz 09:42


    Kyle Adams 09:43

    And obviously, Wall Street’s in New York. Your second point, and this is a big one, we had the New York State Labor Commissioner Roberta Reardon who made this point that AI will enhance and not replace workers.

    Roberta Reardon 09:58

    So let me first say that this is a topic that frightens workers and frightens parents and frightens workers’ representatives, because the fear is that AI will replace people. And you hear that all the time. So I think the first thing we have to do is we have to acknowledge these wonderful examples that you’ve given of how AI actually enhances your life and enhances your workplace and can make your work experience better. But it is a massive change. And this change will happen so much faster than changes ever happened before.

    Laura Schultz 10:38

    One of the speakers on the panel had a really great point about what a doctor’s job will look like. He said that when a doctor reviews medical charts, when you go into a doctor’s appointment, and they review your medical charts, they really only look at the first page, that’s all they have time for. Whereas, AI is going to be able to generate a summary of your entire medical history and everything that might be relevant to why you’re there that day. It’s going to help a doctor look through more information and make sure they have exactly what’s relevant at the tip of their fingers, so that they can come up with a better solution.

    Kyle Adams 11:20

    I think an important underlying point that is no less exciting and doesn’t get as much attention is your third point, those who have the most data win and I think maybe Commissioner Reardon made this point as well, that the collection and management of data is really going to be a major field and opportunity area. So can you explain that a bit?

    Laura Schultz 11:41

    I think we need to view data as a resource. Just like we used to see water or iron. All of those things being really critical to the growth of America, I think data is critical for the growth of this next technological revolution. AI is being enabled by something that we call machine learning. The way that it works is that computers now have the power to look at millions and millions and millions of pieces of data and look for patterns and understand trends that were never available before. The more data that you have, the more intelligent your computers can be, the better conclusions they can reach, the more accurate trends that they can identify. Now New York is a hub of data. We have more medical records digitized than anywhere else in the country. We have major corporate headquarters and those corporations control large amounts of data. Having access to that data means that New York is going to be able to access and create more intelligent artificial intelligence.

    Kyle Adams 13:01

    Professor Jonathan Manes of the University of Buffalo’s School of Law made the point during the panel discussion that we need to be careful to root out biases in this data that already exists.

    Jonathan Manes 13:12

    The vision here is that we might be able to have more objective unbiased systems that we can eliminate the human frailties from these systems. That’s the vision. Increasingly, though, I think researchers in this field have realized that these systems can encode or incorporate the biases that already exists in the data that is fed into the special machine-learning algorithms. For example, if you think about a system that’s predicting where crime is going to occur and one of the major inputs is where arrests have taken place in the past, if the police are focusing on a particular neighborhood or a particular market as community, they’re going to be overrepresented in the data. The algorithm is going to suggest the police go to those places, even if in fact, crime is occurring just as often elsewhere. We have to be aware of these risks that we’re encoding bias, might be laundered biases through algorithms. Sometimes they’ve even exacerbated them.

    Kyle Adams 14:13

    Could you take us through the risk of that a little bit and how we might try to mitigate it?

    Laura Schultz 14:19

    Well, artificial intelligence technologies are ultimately developed by people. It is possible that the flaws of the programmers and the data scientists are going to find their way into the data that gets entered into the algorithms and the algorithms themselves. Developers of this and watchers of artificial intelligence are nervous about the potential implications of these. Also companies are concerned about potentially adopting AI that they need to understand what’s going on in the new technologies that they’re using. Transparency has been an important part of the development of artificial intelligence and its eventual adoption. Companies like IBM who are developing new AI technologies are promoting the fact that their programs are transparent and are giving insights into how the algorithms work so that people understand what the outputs of them are.

    Kyle Adams 15:18

    I think transparency is an important point here that I’d like to expand a little. The panelists spoke multiple times about the idea of AI being a black box or these algorithms being a black box, but I’m not sure we ever got a good explanation of what that meant. Can you tell us what that is?

    Laura Schultz 15:36

    Consider the example of a mortgage application. An AI program can take your entire work history for a decade, look at your tax returns, collect thousands of points of information about somebody’s financial history, and then come up with a yes or no, approve/disapprove. Transparency would require that program to say, approve or disapprove and then list the top five leading factors as to why that decision came about. Then that would give you a better understanding into the factors that made the program’s decision.

    Kyle Adams 16:19

    What skills will workers need? This obviously was a pretty big topic as we talk about transitioning to the world of AI. What were your takeaways from that part of the discussion?

    Laura Schultz 16:31

    The big takeaway is that workers are going to need skills for lifelong learning. The technologies that we work with on a day-to-day basis are going to change dramatically over a decade. You’re going to have to be constantly retrained and learning about these new technologies and updated on your skill set. We need to think about our education system in the K-12 and higher ed. Instead of teaching students facts, we need to teach them how to constantly collect and process new information and process new skills, so that they’re always ready to be learning and updating their skill sets.

    Kyle Adams 17:13

    There was some discussion during that section about the role of the social sciences and what we may think of as software fields in this new world full of AI. What do you take away from that idea?

    Laura Schultz 17:29

    These technologies, AI technologies are going to be more integrated in our day-to-day tasks than anything we’ve really encountered before. Now we go to the device, we go to our computer, turn it on and complete the task using a software. I think AI is going to just be more integrated, constantly collecting information, processing things for us, giving us inputs into that. To make that possible, we need the social scientists to understand how people interact with technology. We need to have a new technology that’s going to be useful for a doctor, for a teacher. We need to understand how they already do the work and create a technology around that that will seamlessly integrate into work. Social scientists understand how people interact with their environment. There’s going to be a huge demand for the people with that expertise to help develop truly effective technology.

    Kyle Adams 18:34

    To continue this line a little bit. There’s so much focus on STEM education now across the country and in New York. Is there a role for arts and humanities in this tech driven AI world?

    Laura Schultz 18:46

    Absolutely, I think that design is going to be really critical for this new technology. I also think that arts and humanities are going to be one of the first real applications of AI. I think gaming, which brings in technology and arts and music and story are really going to be the people who advance AI for a lot of users. I think we’re going to see AI integrated into the arts as well in ways that we can’t even envision.

    Kyle Adams 19:19

    We’ve already jumped kind of into the next point, which is education in the age of AI. One of the points that came up a couple of times, I think maybe first by Commissioner Reardon, was the way education and industry will have to partner to make all this happen.

    Laura Schultz 19:35

    With workers having to constantly learn new technologies and be trained, I think companies need to start seeing workers as a really valuable asset. Without workers and their creativity and their ability to analyze situations, AI is just not going to be effective. I think companies need to start thinking about how to retool their workers, just like they had to retool their factories. Every few years, the factories would shut down for a week, update the equipment and retool. It is reasonable to expect that in the near future, companies are going to have to invest in the retooling of the skills that their workers have. It can be a lot more efficient to send a worker to training for a week versus finding a new worker to do a task.

    Kyle Adams 20:29

    A question that was asked afterward, which was, who is going to be responsible, who’s going to pay for this lifelong learning? It can sound very intimidating to anyone really. A young person entering the workforce or a person in the middle of their career thinking now I have to totally change things up and constantly be taking night classes? It can be a scary thought. What can you say, I guess, maybe ease that fear? Or maybe that’s just true.

    Laura Schultz 20:56

    I think that companies are going to have to take a more active part in their training of their employees. I think that they’re going to have to give the employees the time to go and get the training. They’re going to have to provide the opportunities and the tools for those trainings. It may make sense for companies to partner with local colleges, community colleges, to develop curriculums and to provide the training that their workers are going to need. I also think it’s important to remember that the technology behind education is going to change dramatically over the next decade. It won’t be taking a night off going and sitting in a classroom, it may very well be spending an hour watching a virtual class. Working in New York and participating in a class in California.

    Kyle Adams 21:51

    A question I wanted to ruminate on a little bit is in the history of education, we definitely broke away from and look back critically on a model of schools as factory preparation. That you’re just churning out factory workers. When I hear all of this talk about how Amazon will be able to tell us what they need in a worker and our schools can do it that seems very much to me just like the next generation of that exact same concept. I’m not sure I really have a question here, maybe how is what we’re doing or thinking of doing today different from that model that we criticize from maybe 100 years ago just because they’re more high tech factories.

    Laura Schultz 22:29

    I think when Amazon and other companies work with local education and talk about what they need, I don’t think it’s necessarily Amazon saying we are going to employ 100 employees every year who have knowledge of programming language. I’ve worked with companies and tried to develop curriculum that will be useful for my students in the workforce. It doesn’t change the fundamentals of what we teach. I don’t think that Amazon is going to say we need these people who are very good at sorting, or know how to work these robots, or work with these machines that are going to be talking about high level skills, team building, interdisciplinary work, general understanding of logic and programming.

    Kyle Adams 23:19

    So you’re not concerned the Amazon’s going to be writing the rules, writing the curriculum, students on demand?

    Laura Schultz 23:25

    No, but I do think what they can do is work with local schools, bring their workers in, serve as mentors, and help students see how critical these skills are to a successful career.

    Kyle Adams 23:42

    The final point you make is that New York is getting ready for AI. What is New York doing?

    Laura Schultz 23:48

    New York has a decade’s long investment in a lot of the advanced technologies that are going to be critical for AI. We need quick processors. We need servers to handle all of this data. AI also needs sensors and ways to collect data from the environment that it can then process. These are all technologies being invested in by New York, in our colleges and universities through SUNY, and also in companies that are locating in New York to develop these technologies. Then, we have with the financial services sector, with the information services, with the data sectors, a lot of our companies in New York and New York City are developing the software and methods of AI. I think New York is well positioned to take advantage of the developments in artificial intelligence coming down the line.

    Kyle Adams 24:49

    We’ve heard all of these panelists speak about how we can transition to this world full of AI. A lot of this discussion seems to be about holding our spot in the workforce with machines challenging that spot. That seems to be the conflict that’s being discussed. One point made by Ajay was the need for lifelong learning. He said that the purpose of this would be to make yourself more useful to this changing world. If we’re getting to a point that machines are so efficient that they can replace us on the metaphorical assembly line. Why are we fighting so hard to hold our spot on the assembly line? Why not sit back, relax, and let machines do the work?

    Laura Schultz 25:31

    I agree with that. I welcome AI into my daily work life. I see a day when someone emails me and asks if I want to go and speak at a conference in Chicago. After I say yes, it immediately pops up a little bot that says you’re going to Chicago on these dates, how about this airplane? How about this flight? How about this hotel? And just automatically plans it for me? Or maybe when it’s time to do my employee reviews, it searches through my emails with my employees and generates all of the accomplishments, all of the things that they’ve done in a year. That I’m happy to surrender the minutia tasks to a robot or to an AI device. That way I get to spend my time doing economic policy analysis.

    Kyle Adams 26:29

    Do you see a potential future where fewer people just need to work or everyone needs to work fewer hours per week, or for a shorter time before retiring, where we really reduce the amount of work we do as part of our normal daily lives?

    Laura Schultz 26:45

    I look back on the steam revolution that enabled new factories. I look back at the electric revolution, which changed work. I look at the IT revolution. There are not fewer people employed. We have created more jobs with every single one of those technologies entering into the workplace. I do think that the nature of work will change and maybe the types of hours that we work will change a little bit but I think that any new technology like AI is just going to generate more jobs to be done.

    Kyle Adams 27:22

    Thank you to Laura Schultz, Rockefeller Institute’s senior economist and director of fiscal analysis for joining us today to break down the future of AI in New York. We’ll continue to study and track this issue as we go, so stay tuned to For more reports and analysis on artificial intelligence, you can also find a full video of our AI forum with SUNY on our web site. Be sure to follow us on social media, on Facebook, Twitter, and Instagram @RockefellerInst. I’m Kyle Adams. Thank you for listening.

Policy Outsider

Policy Outsider” from the Rockefeller Institute of Government takes you outside the halls of power to understand how decisions of law and policy shape our everyday lives.

Listen to a full episode archive on Anchor, or subscribe on your preferred podcast platform.