The economic problem AI can’t solve on its own

Artificial intelligence can increase productivity and profits, but technological progress alone cannot guarantee jobs or full employment when there is not enough demand in the economy, writes Graham White.

BRITISH ECONOMIST David Ricardo written by him Principles of Political EconomyPublished in 1821:

‘I believe that the substitution of machines for human labor is often very detrimental to the interests of the working class.’

The question Ricardo raises here is older than mainstream economics.

In a way, artificial intelligence (AI) is the newest context for this problem; However, this is not to deny the importance of concerns about the employment impact of AI or its current role in the fields of medicine and science.

My concern here is the objective economic impact of technical progress, especially the employment effects.

For many, this is the economic and social issue associated with technical progress, whether it is the introduction of the steam engine or artificial intelligence. For many opponents in the economics profession, Ricardo’s instincts were a legitimate concern, even if his logic was unimpressive.

What is technical progress?

For this discussion, it will be sufficient to interpret this as the implementation of new production methods in which at least some inputs are used less and other inputs are not used more.

Interestingly, the major debates about technical progress in the economics profession in the last century revolved around the consequences of the distribution of income, particularly its distribution between wages and profits.

Technical progress means that the economy can potentially pay a higher real wage or a higher rate of return (profit) on output, or both.

Technical progress therefore potentially allows the size of the pie to increase, so to speak. How a capitalist economy would distribute the extra slice of the pie between labor and capital was a much more complex and controversial question.

The flip side, and the more worrying aspect, is that traditionally technical progress in capitalist economies has been predominantly what economists call “labor saving.” It takes less labor than before to produce the same output. If the amount of output to be produced does not increase, labor employment will decrease and unemployment will increase.

Technical progress and unemployment

The question is whether these unemployment effects are inevitable and long-lasting. In other words, the issue is whether the “internal” dynamics of capitalism can be relied upon to prevent such negative outcomes.

The mainstream and dominant response from the profession to this has typically been that “market” mechanisms exist to offset this potential increase in unemployment. If real wages and prices are sufficiently flexible, this will allow labor to become more competitive, thus at least partially offsetting the labour-displacing effects of technical change.

In addition, as the cost of producing goods and services with new technology and the prices of these goods decrease, the demand for these goods and therefore for labor will also increase. (A good article on these arguments Cesaratto et al inside Examination of Political Economy2003).

Such arguments are, in fact, central to the economics profession’s long-held belief in the innate ability of capitalist economies to approach full employment.

This argument, or variations on it, is certain to be put forward in response to concerns about AI, particularly its consequences for unemployment.

For example, recently the chief economist Communication Bank, Luke Yeaman, wrote:

‘Higher productivity does not necessarily mean fewer workers. Often this means employees doing more valuable work. Artificial intelligence is likely to augment, not replace, most jobs, especially by taking over routine tasks. Early evidence shows this is already saving time, freeing up capacity and allowing firms to do more with the same workforce.’

Pay attention to the words ‘to release capacity’. New technology will allow the company to produce more with the same or less.

And so the argument goes: Because more output can be produced with extra capacity, workers who would otherwise be laid off can stay employed. Alternatively, for the economy as a whole, the structural changes brought about by technical progress facilitate new ways of production that provide employment opportunities for those negatively affected by the introduction of new technology.

Where is the extra demand?

What is assumed but left unsaid in this view is that there will be extra demand for goods and services in the economy, and that this demand will support greater output, which will supposedly counteract the labour-displacing effects of technical progress.

Despite their popularity and dominance, these arguments have never been particularly strong, as debates within the profession throughout the twentieth century have shown. Markets left to their own devices, no matter how free, can meet the demand for full employment as follows: Keynes What is suspected (and these arguments are confirmed) is not a correct inference from sound economic reasoning.

An alternative position is available. This alternative would approach the question we started with by asking how influential demand (in economists’ terms) is and how, as a result, production and labor employment can accelerate sufficiently to offset the tendency of new technology to require less labor.

This brings us to the sources of effective demand: household consumption, investment, government spending and net exports.

Without getting too technical, it is problematic to assume that technical progress will accelerate consumption and/or investment sufficiently to mitigate the effects of unemployment.

The consumption impact will depend largely on how the gains from technical progress are distributed between profits and wages.

Regarding investment, installing new capacity requires the expectation of increased demand; Before any investment increase, the necessary acceleration in demand must occur.

Even the notion that innovation itself somehow accelerates overall investment spending is still out on the economics jury.

Interestingly, this leaves only government spending and net exports to increase demand to offset the labor-displacing effects of technical progress.

The net export channel has always been tricky. Suffice it to say, not all economies can experience export-led growth at the same time. Export-led growth benefits only some economies; provided that others are prepared to expand their domestic demand to attract more imports.

What remains is government spending as a source of demand for the economy to provide at least sufficient demand to cope with the technical progress that the private sector is reluctant to make.

Given all the wobbling and heart-wrenching over supposedly excessive public sector spending in this country, this would really set the cat between the pigeons.

Graham White is an associate professor. Faculty of Economics at the University of Sydney. He teaches macroeconomics and history of economic thought. You can follow Graham on Twitter @heterodox_econs.

Support independent journalism Subscribe to IA.