Citrini and the AI doom scenario
A report published last weekend by the obscure financial analyst group, Citrini Research, on the future impact of AI apparently caused a stock market sell-off in software companies. Citrini was little known until its “Global Intelligence Crisis” report suddenly amassed over 22 million views on X alone. The basic message was that, within a very few years ahead, AI ‘agents’ will quickly replace human labour in all sectors of the economy. This would lead to a massive rise in unemployment, followed by a collapse in consumption and a financial crisis in so-called ‘private credit’ and mortgages, thus triggering a recession.
The Citrini authors say that they were not making ‘predictions’ but just setting out a ‘scenario’ that could happen as early June 2028: forecasting a stock market price crash of 38%; an unemployment rate above 10% and a credit and mortgage market meltdown. And all because AI was so successful that AI agents usurped human labour, especially in software and other hi-tech development currently done by skilled tech workers.
How did Citrini justify this doom scenario for the economy, the stock market and millions of mainly skilled workers that so convinced US investors (at least for a day or so)? The main argument was that AI agents developed by the tech giants would be so productive and so effective that companies would make huge profits by replacing costly human labour. But then, said Citrini, millions would have no wages, so they could no longer spend as before, and a consumpton-led slump would be inevitable.
The 2028 scenario was described. “The owners of compute saw their wealth explode as labor costs vanished. Meanwhile, real wage growth collapsed. Despite the administration’s repeated boasts of record productivity, white-collar workers lost jobs to machines and were forced into lower-paying roles.”… The velocity of money flatlined. The human-centric consumer economy, 70% of GDP at the time, withered.” There would be no escape from this doom because there were no countervailing factors to stop it – “no natural brake”. Loss of incomes would lead to mortgage defaults, not by workers with low incomes, but this time by those hi-tech workers who were paid high wages until AI agents took over.
The Citrini scenario dismissed the conventional view of crises as ‘creative destruction’, namely that “technological innovation destroys jobs and then creates even more”. Not this time. Yes, “AI has created new jobs. Prompt engineers. AI safety researchers. Infrastructure technicians. Humans are still in the loop, coordinating at the highest level or directing for taste. For every new role AI created, though, it rendered dozens obsolete. The new roles paid a fraction of what the old ones did.” So the ensuing slump would not correct the crisis because it was not a traditional cyclical recession but a permanently structural one.
That’s because “AI got better and cheaper. Companies laid off workers, then used the savings to buy more AI capability, which let them lay off more workers. Displaced workers spent less. Companies that sell things to consumers sold fewer of them, weakened, and invested more in AI to protect margins. AI got better and cheaper. A feedback loop with no natural brake.” Human intelligence will no longer be needed, because “machine intelligence is now a competent and rapidly improving substitute for human intelligence across a growing range of tasks”.
What are we to make of this doom scenario? Apparently, many investors in the US tech market swallowed it – at least for a day. But they came to their senses when they were reassured by mainstream economists and others that Citrini was laying out a scenario in just two years that was never going to happen. As shown in previous posts,technological innovations take some time to pervade an economy and make a step change in productivity and its impact on the labour force.
The OECD reckons it could take up to 20 years before AI becomes a ‘general purpose technology’ and that assumes that AI models and agents have become experienced and at least as error-free as humans. And a new report argues that it took 100 years to move from Michael Faraday’s and Joseph Henry’s generation of the electric current in the 1830s to electricity boosting productivity growth and transforming the economy. ChatGPT only appeared on the scene five years ago.
For businesses, this means a growing share of businesses won’t have humans at all. They’ll be agents acting on behalf of individuals, interacting with other agents representing sellers, logistics providers and payment processors. A majority of the commercial supply chain could eventually be agent-to-agent.
Or so the story goes – it may not be that simple. There is still a lot wrong with the ability of these agents to talk to each other and provide a reliable service that matches skilled and experienced human labour. Moreover, AI agents are digital, they do not make physical goods, which we still need. To do that, agents will have to combine with robots and that can only be at exorbitant cost of investment. And this is the real scenario for a future recession. Many mainstream commentators on the Citrini paper reckoned that it was ‘pure Marx’ because it posed a consumer collapse without recovery ie. the end of capitalism. But a consumer-led slump and collapse is not Marx’s theory of crises – although most mainstream (and many leftist) economists think it is.
Marx rejected the ‘underconsumption’ theory of crises on many occasions. Marx’s theory was based not on underconsumption, but on overinvestment or accumulation. Capitalist resort to technologies and machines to cut production costs and raise profitablity by shedding labour. But in Marxist theory, only human labour can create value in production, so a contradiction emerges between trying to increase the productivity of labour by doing away with much of it and trying sustain higher profitability. Falling profitability over time leads to falling profits and then an investment ‘strike’ by capitalists. That is the ‘natural brake’ that Citrini claims does not exist with AI. Capitalists stop investing, then lay off workers and it is then that workers cannot sustain consumption. Mainstream critics of Citrini are right in saying that if AI increases productivity so much, it will lead to falling prices so that consumer purchasing power will remain. But they ignore the real doom scenario: rising productivity means less growth in value and eventually falling profitability.
In the 1850s, Marx clarified these two sides of ‘creative destruction’: “As soon as machinery has set free a part of the workers employed in a given branch of industry, the reserve men are also diverted into new channels of employment and become absorbed in other branches; meanwhile the original victims, during the period of transition, for the most part starve and perish.” (Grundrisse). So eventually, new technologies may take an economy forward, but only after a time and at the expense of labour (and not forever).
Mainstream economists suggest that labour could be protected by a tax on AI agents and capital and/or government handouts to the unemployed – these are the usual remedies offered for the Citrini calamity. But that would not be effective if profitability is eventually squeezed. Instead, what is required is the collective ownership of AI technology and its private owners so that any gains in productivity would be used for social needs (reduced hours and increased public goods and services).
There were three reasons why stock market investors panicked on reading the Citrini scenario, despite the holes in its arguments. Investors were already worried about a possible AI bubble bursting if the huge investment in AI models did not deliver sufficient returns. Investors could also see that existing software development companies with human labour are under threat from AI agents; and they also worried that any bursting bubble could spread to the unregulated private credit lenders and cause a systemic crisis. But investors have been reassured by the counter-arguments to the Citrini doom scenario and, for the moment, it is back to business as usual.
No comments:
Post a Comment