Kamal Ghorui
AI to do most white-collar jobs soon? Improbable. But, say, it does. Then even the rich won’t be so rich
A 7,000-word blog post by a tiny New York-based financial research firm spooked US investors on Monday. Titled ‘The 2028 Global Intelligence Crisis’, it paints a scenario in which AI agents have gutted white-collar jobs. Economic mayhem follows. Jobless workers default on loans. A market crash occurs late in 2027, and 57% of S&P’s value is wiped out. By 2028, joblessness in US exceeds 10%, up from 4.3% now. Citrini, the research firm, insists this is only a scenario, not a prediction, but recent headlines about AI agents’ capabilities make it hard to dismiss.
Within a month, we’ve heard the claim that AI can vet legal documents, we’ve seen it render a realistic fight between Brad Pitt and Tom Cruise, and now, apparently, it can bring systems running Cobol up to date. What more will it do by end of this year? It’s threatening to replace writers, translators, artists. Analysing data, diagnosing symptoms, nothing seems beyond its ever-increasing capabilities. That’s why a white-collar extinction event sounds plausible. If Citrini is right, the impending job destruction is best summed up with the film title: Everything Everywhere All At Once.
But is this probable? Perhaps not. Chiefly because AI agents aren’t there yet in terms of capabilities. AI firms have every reason to claim their agents can do this or that real-world task. It convinces investors to keep funding them. But the real world is messy. As IBM clarified – after taking a stock market hit due to Claude’s Cobol claim – “Translating code is one thing. Modernising a platform is something else entirely.” So, AI tools are like bright college graduates who know their subject, but need hands-on experience. In fact, a new study – ‘Agents of Chaos’ – shows when AI agents are left to work unsupervised, there’s a recurring pattern of breakdowns.
The bigger question is: Is an AI takeover of white-collar jobs desirable? In US, only 10% of population earns over $250,000 per year, but accounts for 50% of consumption expenditure. They are all white-collar workers. Make them jobless overnight, and demand collapses. Loan defaults destabilise the financial system. With demand gone, corporate profits plunge, markets tank, and the rich aren’t so rich anymore. Blue-collar workers also suffer when factories cut back output. There’s another downside we must consider. A white-collar office, say Citrini’s team, has 10 or 100 or 10,000 very sharp minds. In an AI world, global business thinking and execution get outsourced to four or five big firms. They can shut down Everything Everywhere All At Once. It’s too dangerous.
END OF ARTICLE
