The future of work under the threat of excessive automation and AI
Abstract:
Wage growth has slowed down, inequality has increased and labor shares have declined in many Western nations over the last four decades. Much of this is related to automation. But automation is nothing new. It has been going on rapidly at least since the onset of the Industrial Revolution in Britain. What has changed in the modern era is not rapid automation, but the absence of other types of technological changes that counterbalance the effects of automation and create new tasks and opportunities for humans. Such counterbalancing technologies have played critical roles in the past, especially for creating opportunities for humans and shared prosperity.
This talk will argue that this technological path was not preordained. It was the decisions of companies, researchers and engineers to double down on automation and move away from other, more human-friendly uses of technology that spawned it. All of this predates AI, but the evidence suggests that we may be pushing AI into the same trajectory, and in the process, squandering its promise and damaging the future of work. If this diagnosis is correct, what is needed is an overhaul of the direction of research, which can only be achieved by major institutional changes and societal pressure on technological companies and researchers.
Bio:
Daron Acemoglu an Institute Professor at MIT and an elected fellow of the National Academy of Sciences, American Philosophical Society, the British Academy of Sciences, the Turkish Academy of Sciences, the American Academy of Arts and Sciences, the Econometric Society, the European Economic Association, and the Society of Labor Economists. He is also a member of the Group of Thirty.
He is the author of five books, including New York Times bestseller Why Nations Fail: Power, Prosperity, and Poverty (joint with James A. Robinson), Introduction to Modern Economic Growth, and The Narrow Corridor: States, Societies, and the Fate of Liberty (with James A. Robinson).
His academic work covers a wide range of areas, including political economy, economic development, economic growth, technological change, inequality, labor economics and economics of networks.
Daron Acemoglu has received the inaugural T. W. Shultz Prize from the University of Chicago in 2004, and the inaugural Sherwin Rosen Award for outstanding contribution to labor economics in 2004, Distinguished Science Award from the Turkish Sciences Association in 2006, the John von Neumann Award, Rajk College, Budapest in 2007, the Carnegie Fellowship in 2017, the Jean-Jacques Laffont Prize in 2018, the Global Economy Prize in 2019, and the CME Mathematical and Statistical Research Institute prize in 2021.
He was awarded the John Bates Clark Medal in 2005, the Erwin Plein Nemmers Prize in 2012, and the 2016 BBVA Frontiers of Knowledge Award.
He holds Honorary Doctorates from the University of Utrecht, the Bosporus University, University of Athens, Bilkent University, the University of Bath, Ecole Normale Superieure, Saclay Paris, and the London Business School.
Summary:
Economic challenges
Labor’s share of GDP has been dropping for the past several decades
Real wages have been dropping for non-college graduates
Significantly affects their quality of life
Men: no growth for college grads, drops for less than college
Women: no growth for high school grads and dropouts, growth if they have some college education
Trend ongoing since 1980s
Inequality growing
Trend strongest in US but inequality has risen across developed countries
US has fewer labor protections than others
Wages (divided into terciles by income)
Wages of Top tercile have risen a lot
Wages of Middle tercile of workers have dropped
Wages of Bottom tercile have risen a little
Middle class is shrinking, falling into lower class, upper class pulling away
Traditional models of labor economics
Technology improves productivity of labor and capital
Everybody benefits in the aggregate
Improved model:
Model technology impact on a per-task basis
Tasks used to create goods
Different tasks have different degrees of automatability
Will be automated if cost of task > cost of automation
Tasks and goods can be substituted for alternatives
Challenge to model
Technology doesn’t enable automatability of all tasks uniformly
Capital doesn’t support all tasks uniformly
Reality: huge variability across tasks
Two types of automation
Replacing: automation of tasks replaces workers doing them
Augmenting: workers become more productive
New tasks increase demand for labor
Technology types
Brilliant: displace workers and significantly improve productivity
So-so: displace workers and marginally improve productivity
Productivity can be separated into contributions of capital and labor
Data
There are aggregate correlations that show general trend: industries with more automation have less labor share of productivity
The more exposed a local market’s industries are to robotics, the worse their employment and wages
Labor declines are coming from routine jobs
~60-70% of US increase of inequality are due to automation
Rise of AI
Huge increase in AI-related job openings over the past several years
Far more job openings for AI roles are lower-tier establishments
Have more AI-replaceable tasks
Are not growing their employment
So AI employees are automating the others out of their jobs
Historically (1947-1987) task replacement has been balanced with new tasks
Since 1987 displacement has been much faster than creatioof new tasks, especially in manufacturing
This is not a necessary outcome
Automation doesn’t increase productivity all that much (~.6%/year), so not such a great gift to society
US tax code actively promotes automation by taxing labor more than capital
Many tech firms are explicitly focused on automation, which is a choice rather than necessity
Companies explicitly rewarded by shareholders for reducing costs
Government support for blue sky research has dropped, reducing creation of new tasks
We may currently be over-automating based on labor costs being perceived as higher than real