Technology’s Bifurcated Bite

Source details

Technological developments (factory robots, smart home devices, self-driving cars) can raise productivity and living standards but also displace workers.
Historical example: post-1960s United States created millions of net new jobs; unemployment rate cited as about 3.5 percent and the ratio of employment to population at a multi-decade high.
Despite job growth, wage inequality worsened beginning a few years after 1960 and the share of income going to workers fell (Karabarbounis and Neiman 2014).

Technological change that improves the productivity of skilled more than unskilled workers.
Reductions in the cost of capital that complement chiefly skilled labor.
Increased ability of machines to replace workers entirely for particular tasks.
Increased concentration of market power in a few firms as a result of technology.

Katz and Murphy (1992): evolution of relative wages as a race between demand for and supply of skilled workers; supply increases can reduce the skill premium, persistent demand increases raise it.
Capital complementarity: machinery and equipment tend to complement skilled workers and substitute for unskilled workers; Berg, Buffie, and Zanna (2018) treat AI and robots as a new type of capital that substitutes for some groups of workers and complements others.
Over the past 30 years, substitutability between ICT (proxy for new technologies, including computers and early AI) and unskilled workers seems to have increased.
Acemoglu and Restrepo (2020): automation of routine tasks explains most changes in relative wages; exposure of labor groups to automation is central.
Corporate market power: firms with access to big data and costly AI development (example firms named in text) can capture greater income from AI, rent technologies to other industries, and potentially reduce labor’s share of income.
Firms may buy and bury competitors, stifling competition and innovation, and can shape regulation to align with corporate interests (Acemoglu and Restrepo 2022).

Rapid evolution of AI (example: ChatGPT-4 emergence in early 2023) marks an acceleration in pace of change and capability beyond routine tasks.
Expert expectations revised: earlier McKinsey-surveyed experts expected computers to write at the level of the top 25 percent of humans by 2050 and perform human-level creative tasks by 2055; those estimates were revised to 2024 and 2028, respectively.
One estimate: once GPT is introduced into the work environment, about 20 percent of workers could see at least half of their tasks affected.
GPT appears to increase productivity in creative tasks (writing, legal analysis, programming); studies find big jumps in productivity for groups using GPT versus controls.
Observations include that the least-skilled participants may benefit most from GPT and that GPT-4 alone may exceed human-level output in some cases.
These findings contrast with prior emphasis on automation of routine tasks and substitution of unskilled labor, suggesting a potentially different distributional impact.

Past research (Alonso and others 2022) suggested substitution of AI for unskilled labor could widen global income disparities, disadvantaging lower-income countries.
Generative AI introduces uncertainty: developing economies could benefit via tutoring and programming assistance, but limited access to data, expertise, and technology could widen divergence.
The prospect of artificial general intelligence (AGI) adds further uncertainty; AGI would presumably be capable of any human intellectual effort.

Higher substitutability of workers with machines and AI can increase wage inequality and the share of income going to owners of capital.
Long-term society may gain from higher productivity, but many—particularly those already less well off—could be losers, and transitions could last decades or be effectively permanent for some workers (Berg, Buffie, and Zanna 2018).
Policymakers must understand distributional implications and prepare for economic, social, and political upheaval under optimistic and pessimistic AI scenarios alike.
Policy questions highlighted implicitly include: how to distribute AI benefits (who owns AI), how to manage corporate market power, how to support affected workers during transitions, and how to ensure access to data and expertise globally.

The article references multiple academic works and surveys to support its analysis and places recent AI advances in the context of historical technological transitions and their unequal effects.

Authors: ANDREW BERG; Chris Papageorgiou; MARYAM VAZIRI — Technology’s Bifurcated Bite, F&D Magazine, December 2023.

F&D: Technology’S Bifurcated Bite
- F&D: Technology’S Bifurcated Bite (Markdown version)
- F&D: Technology’S Bifurcated Bite (PDF)