Artificial Intelligence, Part I – Setting the Stage
Technological disruption is not new. After two industrial revolutions, the third one was about technology and digitization of society. And we are entering the fourth one, inspired by the arrival of “cyber-physical systems” according to the World Economic Forum. As most headlines focus on current applications of artificial intelligence (in its many forms) and dreams of AI-driven future, few tend to debate what policy adjustments are necessary to benefit from the rise of the machines. Depending on the methodology used, most commonly cited studies forecast that AI can potentially replace between 9% and 47% of all jobs in the next decade or two. Carl Frey and Michael Osbourne , using an occupation-based approach, estimate that 47% of all jobs in the US could be replaced. A study by OECD (Arntz, Gregory and Zierahn, 2016) , using a task-based approach, estimate that only 9% of jobs across 21 OECD countries face a high risk of automation. While these studies estimate only potential job losses to AI, meaning none might materialize, the scale of the issue should be obvious.
Despite that, there is little policy debate on the subject. Andrew Yang, running for the Democratic nomination, did elevate the issue but it remains outside the political mainstream. Solutions offered to date, be it Universal Basic Income (UBI) or government facilitated retraining programs, are inadequate. “Artificial Intelligence, Automation, and the Economy” report, published by the Obama administration at the end of his presidency, highlighted several policy responses. To sum up, the administration suggested 1) increased investment in AI research, 2) increased focus on education and training of the new generation, 3) strengthening the social safety net (through existing programs) and re-training the workforce. Assuming the best-case scenario where AI replaces 9% of jobs, as suggested by the OECD report, these policy recommendations might suffice. Taking such a wishful approach, however, is irresponsible.
The potential scale of the AI-driven job displacement and its social and economic consequences for most of us necessitates thorough public debate and increased policy focus. Combined with existing ailments, AI-driven changes to the labour market might be too much for the social fabric of our society.
Focusing primarily on the US, this AI series lays out some arguments for a more pessimistic view on the human prospects in the AI-augmented labour market. On the policy front, from the Green New Deal to Andrew Yang’s UBI and social credits proposal, there is a growing push for more radical policy solutions. I certainly agree with the notion that significant reforms are necessary to prepare us for the reality of an AI-driven world. However, in the age of constant political campaigning and social media, many proposed solutions currently discussed are not grounded in reality. The Green New Deal, for example, championed by Alexandria Ocasio-Cortez and some 2020 Democratic candidates, urges the government to adjust its environmental policies, invest in crumbling infrastructure and uplift the economy in a just and fair way. But it does not acknowledge some of the most basic failures of American democracy and capitalism. We should address the structural issues first, before pushing for progressive reforms on climate change and AI-driven job displacement.
Part II will paint a narrative about the current state of the American economy, pre coronavirus disruption, and discuss why the US is fundamentally unprepared to deal with the fourth industrial revolution.
 Carl Frey and Michael Osborne, “The Future of Employment: How Susceptible are Jobs to Computerization,” Oxford University, 2013, (http://www.oxfordmartin.ox.ac.uk/downloads/academic/The_Future_of_Employment.pdf)
 Arntz, M., T. Gregory and U. Zierahn (2016), “The Risk of Automation for Jobs in OECD Countries: A Comparative Analysis”, OECD Social, Employment and Migration Working Papers, No. 189, OECD Publishing, Paris. http://dx.doi.org/10.1787/5jlz9h56dvq7-en