Forecast for 2028: AI Gets Smarter, Crisis and Unemployment on the Rise
AI technology development does not always end well. In fact, not long from now, AI could cast the world into gloom by 2028, with an economic crisis, scarce jobs and rising unemployment.
The 2028 crisis forecast due to AI is not mere fantasy. It is an thought experiment prepared by investment research firm Citrini Research together with analyst Alap Shah in a report titled ‘The 2028 Global Intelligence Crisis’.
From the outset, Citrini Research said that their report is not a definite prediction, much less an exaggerated doomsday narrative.
Their report is a thought experiment, a simulation of extreme risks if AI truly surpasses expectations and displaces humans too quickly for the economy to adapt.
So how could AI’s success become a boomerang that destroys the world economy? Here is the anatomy of the crisis, as summarised by KompasTekno from Citrini Research’s site.
It all begins in a golden era filled with blind euphoria. By 2026, the world economy appeared to be at the pinnacle of its prosperity thanks to AI that was becoming ever more advanced and efficient.
Positive sentiment towards AI drove shares higher for an extended period, and the technology sector was the main driver. The mood of optimism was very palpable. Investors were highly enthused; markets were full of confidence.
By October 2026, the US stock market was partying. The US stock indices surged to levels previously hard to imagine.
The S&P 500, the index of 500 large US companies, approached 8,000. Meanwhile the Nasdaq rose smoothly to breach the 30,000 mark.
At the same time, the first wave of layoffs began in early 2026.
Many companies cut jobs as some functions were deemed replaceable by AI systems and automation. The term that circulated at the time sounded cold: human obsolescence, in which humans were becoming increasingly obsolete in several lines of work.
The role of office workers began to be massively replaced by AI agents that were overly efficient.
For shareholders, this phenomenon was a priceless boon. The business logic appeared to operate perfectly on paper. Mass layoffs meant radical reductions in operating costs.
What happened next? Profit margins widened rapidly, earnings reports consistently exceeded expectations, and stock prices soared to the skies.
However, trillions of dollars of record profits were not used to create new jobs for people.
Instead, those colossal funds were recycled and injected into buying more AI computing power, more GPUs, and more data centre infrastructure.
This cycle created the illusion that the economy was running faster, even as the fundamental underpinnings (real human consumption) began to crumble.
On paper, nominal GDP numbers did grow rapidly.
AI-based companies saw their wealth explode when labour costs vanished. Productivity also surged. Real output per hour rose at a pace not seen since the 1950s.
All of this was supported by AI agents that never sleep, never take leave, never get sick, and never claim health insurance.
Yet behind this ‘too good to be true’ situation lay one fundamental flaw that was overlooked. Machines do not shop. AI would not spend on essentials or discretionary consumption.
Essential or basic needs include staples such as rice, electricity, water, and school fees.
Discretionary consumption, meanwhile, is spending on non-essential or non-urgent goods, such as holidays, concerts, buying a smartphone when the old one still works, dining at expensive restaurants, and purchasing luxury items (bags, clothes, watches).
The bitter truth is that 70 per cent of the economy’s support comes from human consumption. When a cluster of GPUs in a data centre can produce output equivalent to 10,000 office workers, that is not a cure for the economy but a pandemic.
At this moment, economists began popularising the term ‘Ghost GDP’. This is the phenomenon where production output is recorded high in national accounts, but the money never circulates in the real economy.