Luohan Academy

Microsoft’s Practices in Narrowing the Opportunity Gap by Leveraging Digital Technology

Event materials

  • Transcript


Mike Schwarz

Thank you. It's a pleasure to be here and it's a pleasure to be part of this community. And that brings me to the first observation about what we mean today by inequality. We know that inequality within most developing countries is increasing. We also know that the inequality on the globe overall is actually shrinking, primarily because of the rapid economic development in India and China that is narrowing the gap with the wealthiest economies. India and China are very large share of world population. So when we think about inequality and think about those trends, let's say nobody's incomes actually change. Let’s imagine for a second that everybody's incomes stay the same, but the notion of communities changed because of technology that we are using that enables teleconferences, because of the technology that allows communication and remote work.

Right now, I'm speaking today from my home in Berkeley, California. Some of you are in China, and some of you are in New York City and some of you may be in Europe. We are thousands of miles away, but we are part of the same community. And very much so, as I recognize faces. I am happy to be here talking to you. So I am surely a member of this community just as much, actually more so than I'm a member of my neighborhood association.

So then the question is, if we are now members of communities, that are not limited even by nation state boundaries, how should we think about a relevant comparison group for ourselves? So even in the world where incomes didn't change once you break the boundaries of communities, how do we feel about inequality? So I don't know the answer to this question. That's one of the problems that technology can throw in front of us, left front and center. So today we live in a relatively prosperous world, there are more people than ever, there is less hunger than ever. Great. But technology changes the way we live. It destroys some jobs, creates new jobs. We are all part of that process. I think when you look at income inequality and technological change in the 20th century, you see that a lot of that is driven by trade. And AI will mention some of that back in the old days with Sears Reebok catalog creates trade where that creates a lot of value, also create a lot of disruption for a lot of businesses.

So we live in a similar era. But now something else and throughout the 20th century you see lots of that. You see manufacturing moving to China, manufacturing moving to Thailand, to India outside of the US. You see many other new jobs emerging, you see winners and losers. But all of that primary interface between countries at that time is through the channel of trade.

So today, I think that entering in a new era where not only goods can be moving through state boundaries, but also labor could be moving through state boundaries without us physically moving anywhere. And that's a new page in the history of the world economy. And that changes, of course, made possible by wonderful communication technologies and information technologies that are coming online. And overall very much agree with Professor Townsend, it's a very positive thing for the mankind in the long run. But in the short run, there be the winners and the losers. 

So that brings me to get the next really important point. Right now as the movement of labor, people work remotely. In some ways, it's a very new thing. We just started doing it at a scale during the pandemic. And of course it was sort of coming for a very long time. Let's see how long ago it started. Well, if you look at the most software companies, people at Microsoft people at Google, people at Facebook and Amazon and Apple, until very recently nobody worked remotely.

So you can say well we just discovered remote work was possible very recently in some ways, it's true. In some ways, it's completely false. Let's look at one of the most influential software projects of the 20th century, early 21st century. I would say that would be the development of Linux. Almost all the work was done remotely. And that was many years ago. So the possibility of remote work was there. But to a very large extent, it just wasn’t work. A lot of Linux development was done by developers who were in some ways, donating their time or someone else was donating their time. So the “work” itself became remote very recently, but in some ways, the possibility of creating software by multiple workers was there for more than 20 years. So moving on, what does remote work means for inequality and what does Microsoft do in terms of that. So let me give you a few examples of what we as a company do to create a more inclusive world. And I think that to a large extent, we can say it's the job of the government to do those things. The job of private corporations is just to profit maximize.

I don't believe that. I think that large companies like us have social responsibility to make a world a better place. And they are doing a lot of that in various dimensions, some of them are very local.

For example, Microsoft contributes $250 million towards creating affordable housing in Seattle, next to our headquarters. That's where we are. That's our home. And we are doing a lot of things to kind of promote more equitable life there. But then there's a bigger question. So what about the world as a whole? We can’t solve all the world's problem but we can do something. Let me give you a few examples that I'm actually really proud of that Microsoft is doing. First, Linkedin is a part of Microsoft, and like my team does a lot of work in collaboration with Linkedin.

One of the things that Linkedin does that I think people never talk about, but I think it's really important. One of the things for creating kind of harmonious society is having a society where people feel in some ways fulfilled. People don't feel like the anger is amplified. People feel some harmony. When we go to some social network, you see that they grab engagement by promoting things that get people angry. Well, start using Linkedin. Linkedin’s algorithm is deliberately to create a very civilized experience at the cost of engagement. I'm mentioning here are the kind of inclusive society, because really I think that's a big part of it. One of the reasons that our governments often couldn't deliver on creating more inclusive societies, because we can be very politically divided in many ways. We need social platforms that would create cohesion rather than division.

Here's another wonderful thing that Linkedin does that probably most of you never heard about. Linkedin tries to lower digital divide by creating thousands of courses by partnering with other companies. In creating thousands of courses, many of them targeted for people with low skills. And those courses, you can learn anything. You can learn how to video edit, you can learn how to audio edit, you can learn how to do secretarial tasks, you can learn some python and some basic data science. You can learn a lot of things and we are trying to promote those courses. Many of them are very short. They could be half an hour, an hour, 4 hours. They help people to create skills and lower the digital divide, which I think is really important. That brings me to another point about this world without distance, world where remote work is possible.

On one hand, this remote work gives the opportunities to anybody where in India, in West Virginia, in California. And yet you see most remote workers living in very affluent areas and people who in theory have opportunity in areas that may be less developed actually don't take advantage of those richer labor markets. Why? Because they lack skills. That's why kind of we are really trying to partner with both municipal organizations and even other companies to promote that content that could help human development of people who are lower skilled. That's one of the things that we do. Let me mention very briefly, another thing that I’m very proud of that we do. That's our work on AI ethnics. I myself as a member of the committee that kind of tries to create principles for inclusive AI. To tell you the truth early on when I just got involved in that effort, I thought why should we worry about AI algorithms being fair? After all, if there's any unfairness in the marketplace as Gary Baker’s argument would go, there would be an arbitrage opportunity and the market would arbitrage that away. I don't think it actually works like that. Let me give you one example, just an anecdote that convinced me that we really need to worry about AI algorithms being “fair”, one of extremely prominent and extremely smart economic theorist whose name I'm not going to name here, told me that early on in his career, he had a southern accent. He’s white, male, but with a southern accent. He felt that he was extremely penalized by the job market because of that, because people perceived him as stupid. So he actually made an enormous effort to learn how to speak without a southern accent.

And then he had a very successful career and felt it paid wonders for him. Now let's go to AI algorithms. Now we have hundreds of applications for many jobs, and we have more and more reliance on AI algorithms for screening that application, you can have all kinds of biases that could be quite terrible. Let me mention one bias that's relatively innocuous. If you look at AI software, that is trying to recognize images, and you look how well it recognizes huskies from wolves, very different animals, but look very similar. It does pretty good. It's about 85% accuracy. Then look at explainable AI, researchers tried to figure out how that does it. So trick the picture, see what does it. So it turns out that the main feature that the software was using was snow, The way that it knew it's a wolf and not a husky is that there was snow on the background. Remove the snow, it turned out that you train on snow and not on wolves. So think about it. If your software is not explainable, you may end up with massive amount of discrimination in the labor force. Just like my friend with southern accent was discriminated earlier on in his career. And the arbitrage wouldn't really be kicking in that strongly or maybe not at all. Because if you're making lower job offers to people, because you perceive them being inferior in some way, everybody else also makes them lower job offers. So there's not a huge benefit from increasing them. And unless there's a very strong, some kind of performance feedback signal, it will take a generation for those kind of inequalities to go away.

So I think I'm out of time, but I would like to conclude with paraphrasing the, quote from Homer Simpson: “Technology is the cause of and the and solution to all life’s problems.” Thank you.


For more information, please visit Luohan Academy's youtube channel: Luohan Academy


to leave a comment