A lot of the conversation about the future of AI and automation focuses on the AGI endgame (“will humans still work when artificial general intelligence can do everything?”). But there are more interesting, tractable, and concrete questions to answer about the effects of “narrow,” task-specific AI that looks more or less like what we have today. In the near future, we can expect more advanced robotics, autonomous cars, customer service chatbots, and other applications powered by such narrow AI to take over certain tasks from humans. Should we be optimistic about labor in the next 10-50 years, when parts of industries will be automated by narrow AI? What early signs of those trends should we be concerned about now?
There’s little controversy that AI adoption will cause short-term job losses – as has been the case with past industrial revolutions that ultimately left most better off. The question is whether AI-enabled automation will bring about lasting unemployment – is this time going to be different? The commonly held view is no: as we reap the benefits of tech-enabled productivity gains, as with past technological advances new sectors and jobs will recoup the losses. Just imagine a more efficient future of autonomous vehicles, with new jobs around redesigning cities / infrastructure and more time for otherwise frustrated commuters to work and play. In Mary Meeker’s “2018 Internet Trends Report,” (a staple of the startup media diet), she echoes this thought:
“Will technology impact jobs differently this time? Perhaps … but it would be inconsistent with history as … new jobs/services + efficiencies + growth typically created around new technologies.”
As we’ll see, automation might not inherently work this way. The way we deploy AI-driven technology in the near term may change whether we see lasting negative effects on jobs.