Before the breakthrough of digital cameras, a photo could cost a quarter of a dollar or perhaps a bit less, including the new film needed. Today, with digital cameras photos are almost for free. One can make many observations based on this technology shift. One is the fact that very few people have any photos printed nowadays. Another is that, even if there were also those that never sorted or organized their material photos earlier, today, with digital photos, that seems to be the rule. Most people seem to have thousands over thousands photos in their digital storages, but few have any order in it. Also, as photos are so cheap, people take a lot more photos, including twenty shots of the same motif – without ever deleting them. These are interesting shifts in behaviour that can merit some deeper analysis.
But today, I want to use this shift in photography for a discussion about ”productivity” and economic growth. Clearly the ”productivity” in consumer photography* has increased a lot. Today you get many more photos for a very low cost. Admittedly, a paper photo and a digital one is not the same, but they are comparable, and people apparently skip the paper photos voluntarily, after all one can also get paper photos from digital shots. But that is not what economists mean with productivity. For them productivity is measured in money: ”The most common productivity measure is labour productivity, defined as economic output (gross domestic product, or GDP) per hour worked”. So, while ”productivity” in photography from a consumer perspective has improved a lot, from an economic perspective the value created, aka the contribution to GDP, has gone down considerably. There are similar experiences from the music industry where the shift from CDs to streaming meant a considerable reduction in the revenue of the industry. In 2014 the global revenue was US$13.1 billion, $9 billion less than it had been 15 years prior. Even today, despite dramatic increase in streaming, incomes and population, the industry generate less than in the early 2000s, if inflation is taken into account.
According to an article in The Economist, What if AI made the world’s economic growth explode? (24 July 2025), artificial general intelligence (AGI), capable of outperforming most people at most desk jobs, will soon lift annual GDP growth to 20-30% a year, or more. AGI, the theory says, would allow for ”runaway innovation” supercharging growth in GDP per person. The projections are based on that most of the world’s economic output will be automated. Automating production requires only (my italics) sufficient energy and infrastructure—things that more investment can produce. If machines get sufficiently good at replacing people, the only constraint on the accumulation of capital is capital itself. The article claims that goods made in AI-run factories could be close to free and that food prices could ”collapse”.
Let’s deconstruct the scenario in the article. If we start from the end. Things that are ”for free” don’t contribute to the GDP at all. Things that have become very cheap don’t contribute much to the GDP either. Producing more with less input (labour or other) in a sector leads to falling prices and/or increased consumption. In the first case, it will lower GDP, in the second case the fall may be offset by increasing consumption. An highly mechanized farm today produces several hundred times more per worker than a manual farmer. Most of the productivity increase in farming resulted in lower food prices and some of it in higher consumption or that people climb the food quality ladder. As people just can eat so much, a large share of the increased grain yield is converted into meat which, from a farming perspective is about added value. Also, the consumption of resource demanding vegetables and fruits has increased a lot (much more than meat actually).
The tremendous increase in the farm sector has done little to increase the GDP. If you compare the agriculture contribution to the GDP for a group of very poor countries with a group of rich countries you will see that for countries with a GDP per capita of around US$2,000 the agriculture contribution to GDP is in the range of US$400 to US$950 per capita, while for countries with a GDP of US$70,000 the agriculture contribution to GDP is in the range of US$550 (Denmark) to US$1200 (the Netherlands). The contribution to the GDP of the agriculture per capita in the United states is the same as in Afghanistan and considerably less than in Ethiopia (obviously, it is a lot higher if counted per agriculture worker). From this one can get the impression that the enormous productivity increase in the farm sector has not had any impact on growth.
But wait a second. The effect on the GDP is apparently not within the farm sector itself. The impact of the productivity gains in the farm sector is that a lot of people that were made redundant could move to the growing industrial sector. To some extent, there is also a GDP-effect in that food has become cheaper, and that people can work better and live longer if well fed. The industrial sector has, in turn, automated. Here the possibility to increase consumption has been bigger than in the farm sector. But as population growth tapers off it is harder for ”us” to consume all the stuff, and prices of industrial products fall. The share of the industrial sector’s contribution to the GDP also starts to shrink.i The service sector grows, even if some of that growth is just a result of outsourcing jobs that was formerly made within industrial companies.
The only thing that will generate economic growth in the quoted article is AGI making machines for making machines for making machines….The economist apparently believes in some kind of economic perpetuum mobile where there is no labour, no humans and fundamentally no nature – only capital that accumulates more capital (machines are capital). Well, the need for energy is mentioned, in passing. The article comforts us that ”more investment” can produce the energy needed, so even the supply of energy is just a question of capital.
Perhaps Marx, and others before and after him, went a bit far in ascertaining that it was only labour that creates economic value. The opposite notion that it is capital that creates more capital appears to be validated by rents of various sorts. But capitalists claiming rents don’t constitute an economy,as it is only a small, mostly parasitic part, of an economy. I believe that anybody that thinks it through will understand that if there is no labour and no consumers, there can be no capital accumulation in the aggregate. A worker with no income will not pay her mortgage, and nobody will buy the Teslas if there is no income from work. If AGI makes machines making machines at no cost, there will obviously be no money involved and basically no value created at all.
The only way that AGI could spur unprecedented economic growth is if most people made redundant by AGI are occupied in other non-AGI work. It is a bit hard to see what kind of work that could be if AGI will accomplish all that is claimed. For sure, people would find things to do, we are an industrious species. But it is only if that becomes a job with income and thus that there is a market for the stuff or service it provides, that it will contribute to the economy. Ironically, if AI or AGI would deliver on the promises floated in the Economist, it would most certainly mean the end of capitalism, the market and a monetary economy. Meanwhile, if AGI is so smart, it will also do away with the main argument against a planned economy, the deficiency in information. If AGI will deliver it, that will most likely result in some kind of bureaucratic state with no capitalists, a socialist utopia.
In reality, apart from the misunderstanding of what economic growth is, I believe that AI will mean as little for productivity as digitalization has. Humanity has embraced digitalization whole-heartedly and, for sure, there are benefits, such as me publishing this article and you reading it. But by and large digitalization has done very little for growth apart from within the digital sector itself. Personally, I am not at all convinced that “they” will be able to produce such brilliant AGI as the Economist projects. I am not too impressed by what I have seen so far, and apparently most industries (apart from the AI industry itself!) are fairly slow in taking AI into their core business. In my view, AGI should rather be spelled out Artificial General Illusion. Resource constraints might be an insurmountable obstacle for a global AGI revolution, but that is yet another discussion. As I wrote in another recent article, I see AI mostly as a way to cope with ever increasing complexity of modern civilizations, the controls and management needed and the ocean of data that has followed in the wake of digitalization. Once in place, the technology will get uses we couldn’t foresee.
This article is about the effect of AI on economic growth measured as GDP. Some of my readers will perhaps find it pointless to discuss the impact of AI on GDP as GDP is a flawed measure and growth is nothing to strive for. Those who follow me regularly know that: 1. I am more into the no-growth/de-growth camp, 2. I don’t think there is any good link between GDP and human well-being, and 3. I think that GDP as a measure has many flaws. I still believe, however, that the GDP as a measure of economic activity has some relevance and as I am exploring the link between the economy, the human ecology and the resource throughput, I find it interesting to get a grip on it.
* This is not about professional photography. I have no expertise in professional photography, but I assume that the cost for it is mainly about human work and creativity, and whether the camera is digital or not is not so relevant for the economic discussion. There is of course less work with the development of photos, on the other hand, going through ever much bigger numbers of digital pictures and working through the programs to enhance them also takes a lot of time. I assume that the time saved is rather small.
i Marsh, Peter 2010, The New Industrial Revolution, Yale University Press.
Teaser image credit: Why on earth did I take that photo? Gunnar Rundgren.