🐙 After AI, where will the jobs come from?

In my commentary this week, I look at the insightful research note from Goldman Sachs, titled “The Potentially Large Effect of Artificial Intelligence on Economic Growth”, in which the authors explore the labour market's future.¹

The study posits that global productivity could see an impressive uptick, ultimately boosting global GDP by 7%. In the U.S., economic growth could leap from the anaemic range of 1% to 1.5% to a more robust 2.5% to 3%, reminiscent of the booming 1950s and 60s.

The research highlights that AI automation affects two-thirds of U.S. occupations to varying degrees. A significant portion of these jobs could see substantial replacement due to automation. The researchers propose that jobs with 50% or more of their tasks automated are at risk, while those with 10-49% automation will see AI complement human efforts.

Historical evidence suggests that companies adopting new technologies tend to expand their workforce while others lag behind². Firms investing in AI are poised for better performance, but the landscape will inevitably include both winners and losers. Moreover, widespread downward wage pressure is anticipated as workers become more efficient and the demand for human labour diminishes.

The temptation

I wondered about how organisations might respond to this.

Here is one hunch, when thinking about medium and larger firms. Middle-tier employees—well-compensated but not part of top leadership—may find themselves in the crosshairs of cost-cutting measures. Enough of this tier would be retained to guard the tacit knowledge, the accumulated social learning, of how the firm actually works. AI-enabled junior workers could collaborate with this leaner group of seasoned senior professionals to bridge the gap.

I suspect this might be quite a tempting path for top management.³

Where’s the new?

The projected 7% productivity growth would fuel a faster-growing economy. This economy will have new demands. We’ve seen this picture before. Over the past 80 years, a striking 85% of U.S. employment emerged in jobs that didn't exist in 1940.⁴ Innovation works!

It’s always worth bearing Schloss’ Lump of Labour Fallacy in mind. According to GPT4’s output, the fallacy is

the mistaken belief that there is a fixed amount of work available in an economy, and that creating more jobs for one group of people will necessarily take away jobs from others. This fallacy ignores the fact that economies can grow, creating more jobs, and that increased productivity can lead to higher demand for goods and services, generating even more employment opportunities.

However, the swift deployment of AI technologies could lead to job losses outpacing the creation of new roles.

Addressing the challenges of reskilling workers and ensuring the emergence of new jobs in suitable locations requires nuanced and effective policy interventions. Both private and public sector solutions must be devised and implemented at scale in the coming years.

I would love to hear your thoughts: what kind of private or public sector interventions could speed up the creation of new types of work? Where are there good examples which are already having the right effect? Share your ideas in the comments below.

1. Sorry, the paper isn’t available for download.

2. I discuss this in chapter five of my book.

3. One survey from a software firm, suggested 66% of managers say they would gladly replace employees with AI tools if the work was comparable. I caveat this because it is an in-house survey rather than an academically rigorous piece of research.

4. Autor et al, New Frontiers: The Origins and Content of New Work, 1940–2018