The rise of complex algorithms and machine learning plays an outsized role in the public imagination. Just last week, the New York Times Magazine published a long cover story on Google Translate’s (machine-learning-based) leaps forward. We are all watching with great interest as driverless cars take to the streets. And commentators increasingly focus on the possibility that machine learning will displace a significant number of jobs as it becomes embedded in sectors across the economy.
Up to now, the conversation about job displacement has been separate from another important technology-related discussion: the role of government investment in technological innovation. But the future of the social safety net may depend on bringing these two conversations together.
As we contemplate the potential necessity of a much more robust social safety net—a necessity brought on by technological change — we should also be talking about whether we, the taxpayers, are getting a reasonable return on our tech investment tax dollars. This connection is crucial, given that the extent to which technology displaces jobs will probably be closely tied to the extent to which government technology investments bear economic fruit. And retaining an equity stake in the technology we develop could mitigate the need to raise taxes—imperfect means of recapturing value that they are, with any number of loopholes and other limitations—on machine-learning-based profits after the fact.
The Possibility of Significant Job Displacement
We don’t yet know whether machine learning will fundamentally change the demand for labor. Some argue that the economy will adjust to technological change just like it has for 200 years, and jobs (or the demand for work) will be created at something approximating the rate at which they are displaced. Those making this case sometimes point to the fact that we would expect to see productivity increase at a higher rate than it has if we were heading for significant displacement.
Others argue that we are approaching an inflection point when the economic relevance of technologies related to machine learning will take off and, given the pace of the coming change, the labor market will not be able to adapt, at least not in a timely way. The seemingly constant arrival more quickly than expected of machine-learning milestones might bolster this argument.
In Either Case, Better to Be Safe Than Sorry
But even if it’s not yet clear which scenario is more likely, the possibility of unprecedented job displacement makes the notion that we will need a more robust social safety net in the not-too-distant future worth thinking about. Some, like Andy Stern in his recent book, Raising the Floor, have argued that this social safety net should take the form of a universal basic income (UBI) (another topic of the moment). While that may or may not be the answer, the amount of money required for a UBI does serve as a wake-up call: whatever the makeup of the safety net in a high-job-displacement scenario, it will not be cheap. We need to be thinking about how we would fund it, and that brings us back to a reasonable return on our collective tech investment.
Mariana Mazzucato has written extensively about the role of government investment in technological innovation. And it’s enormous, at nearly all stages of development. As Time’s piece on Mazzucato’s recent book, The Entrepreneurial State, pointed out: “The parts of the smart phone that make it smart — GPS, touch screens, the Internet — were advanced by the Defense Department. Tesla’s battery technologies and solar panels came out of a grant from the U.S. Department of Energy.”
The government invests in research and development for many reasons. National security. Public health. Energy needs. Space exploration. Economic growth. Most would agree that these are important investments in our collective success. But should we be getting more in return than we currently do, particularly when that same technology may be the cause of strains on the social safety net?
The government’s tech investments have generally taken the form of in-house research, grants, or long-term committed financing through loans. Mazzucato has suggested that the government should consider retaining a financial stake in these investments for the purpose of creating ongoing funds for continued research. That may well be a wise use of the return on our public tech investments. However, the potential for significant job displacement presents a different (though not mutually exclusive) reason to take a hard look at retaining a financial interest in our tech investments, and to actively explore the best ways to structure such an equity interest.
It is time for us, as a nation, to think about keeping an economic interest in the technology we develop, so that everyone who pays into the initial investment can own a stake in technological progress. If machine learning does displace work at significant levels, it will coincide with a widespread adoption of technology in which the government was a “seed round” investor. And if we act quickly enough, the payoff from that technology can play a major role in funding the next generation’s social safety net, whatever form it may take.
Note: The views expressed here are those of the author and not of the United States government.