AI Is The Way Out Of Low Growth And Inflation

Curbing inequality will depend on sharing the new wealth and upskilling labor to co-work with intelligent machines.

Ibrahim Rayintakath for Noema Magazine

Nathan Gardels is the editor-in-chief of Noema Magazine.

The low growth, declining productivity and persistent inflation afflicting the global economy in recent years is about to get a transformative booster shot with the advent of generative artificial intelligence.

That is the hopeful prognostication of former U.K. Prime Minister Gordon Brown, Nobel economist Michael Spence and venerable investor Mohamed El-Erian in their new book, “Permacrisis: A Plan To Fix A Fractured World,” written with Reid Lidow.

There is little doubt in the minds of these authors that the productivity leaps unleashed by AI will generate vast new wealth by broadly expanding the provision of goods and services. The question is whether such a high-growth trajectory will further exacerbate inequality or lessen it. And that largely depends on whether growth is inclusive and the rollout of robots creates more new occupations for gainful employment than it displaces jobs and depresses wages.

Signs point both ways on the impact of AI.

In one study earlier this year, Goldman Sachs predicted that the enhanced productivity of generative AI would raise global GDP by 7% over the next decade. At the same time, another study by the investment bank warned that 300 million full-time jobs could be automated by platforms like ChatGPT. The impact would be mostly on white-collar administrative labor in the advanced service economies, less so in physical occupations like repair work.

Productivity Is Not Where The Jobs Are

As Brown and company point out, the core problem behind the protracted slump has been that productivity growth is taking place where most jobs aren’t — and where most jobs are, productivity growth has stalled.

In the U.S., “productivity is higher in tradable sectors overall, and more importantly, productivity is growing in the tradable sectors much faster than in the non-tradable sectors. In fact, the non-tradable sectors have been in a productivity backwater for more than 20 years. This rut is significant because the non-tradable sector is a huge part of the economy,” they write.

“The tradable economy,” which includes knowledge work and digitally enhanced manufacturing, “accounts for one-third of the overall economy. The non-tradable sectors combined — think government, healthcare, hospitality, retail, education and construction — account for nearly 80% of total employment and the remaining two-thirds of the economy. And it is in this non-tradable sector where productivity is most lagging.”

That gap, they warn, is a “prescription for having a dual economy marked by the have-nots and the have-a-lots.” The best chance to reverse this pattern, they argue, is “the expansion of the digital footprint” to those least productive sectors of the service economy.

Automate Or Augment?

Certainly, digitizing the service sector, as we all know from the personal experience of automated bank tellers to Amazon orders that arrive the next day, improves the productive use of time. In the first instance, it definitively eliminates jobs. In the second case, brick-and-mortar retail jobs are displaced while new tasks are indeed created, mostly for low-wage warehouse workers along with those who physically transport and deliver the goods.

In the manufacturing sector, the point of digitization is both to reduce or eliminate labor costs and improve efficiency. If “innovations within innovation” follow the pattern of previous technological diffusions, they will serve to augment the efficient performance of tasks while also creating entirely new occupations, for example software engineers who must program the robots for ever-more diverse tasks.

The scale and scope of all this is speculative and undetermined since we can’t know what will emerge, how, when and where as AI courses through the economy. But two things are certain: First, to the extent that AI divorces employment and income from productivity growth and wealth creation, the inequality gap will only widen if the new wealth generated by intelligent machines that displace gainful employment is not widely shared but accrues narrowly to a plutocracy that owns the robots.

Second, if AI is to augment instead of fully automate labor, the workforce will have to be upskilled to manage machines more intelligent than we are.

Pre-Distribution And Public Higher Education

The best way to remedy widening inequality in the digital age is to spread the equity around. That entails fostering ways for people to share the wealth created in the first place as high growth takes off instead of trying to mend the inequality gap after the fact through redistribution of income.

We call this “pre-distribution” through “universal basic capital.” Personal accounts could be assigned in mutual-fund-like universal savings plans that are invested in the shares of a broad array of tech companies, including those that profit from AI-enhanced productivity at the expense of jobs with livable wages.

As the economic pie grows larger through AI improvements in the service sector, that value added needs to be properly taxed to finance public higher education institutions that will be primarily responsible for upskilling a workforce so it will be knowledgeably aligned with the new technologies.

As it stands now, public higher education in the U.S. is underfunded. For example, the California State University system, with 480,000 students and 23 campuses, is widely regarded as the surest route to upward mobility in the Golden State. But it suffers an ongoing $1.5 billion operating deficit to meet the heavy demand of the state’s young population, forcing it to close the gap by raising tuition to a nearly unaffordable level for its mainly Latino and Asian students.

The very services where productivity has lagged, but will improve through the diffusion of digitization, account for 70% of California’s $3.6 trillion economy. Yet, aside from odd items like gift wrapping or welding, the service economy is not taxed. As calculated by the Think Long Committee for California, a 1% sales tax on business-to-business services (such as accounting and financial or legal consulting) would generate $7 billion in revenue for the country’s largest four-year university. It is a virtuous circle of capturing new growth to invest in upskilling labor that will spur and sustain further growth.

The integration of AI can jumpstart slumping economies and lagging sectors and set them on a high-growth trajectory as Brown, El-Erian and Spence argue. But for those gains to be fairly shared, the counterpart of innovation must be robust public policies that ensure there will be more winners than losers in the arriving digital age.