Citrini’s AI Apocalypse Report Met with Skepticism and Rebuttals
On February 23rd， a smaller research firm， Citrini Research， released a report titled “The Global Intelligence Crisis of 2028.” This report presented a hypothetical scenario of an AI-driven economic crisis in 2028， including predictions of soaring unemployment and consumer collapse.
The report’s claims included: “US unemployment surpasses 10%， consumption collapses， and participants of the ‘Occupy Silicon Valley’ movement will camp outside OpenAI and Anthropic offices.”
The timing of this report was unfortunate. It coincided with Anthropic’s release of AI plugins for various industries and a blog post stating that AI could write COBOL code， which led to a significant drop in IBM’s stock price.
A confluence of factors created panic in the market， leading to a bloodbath in US stocks on Monday: “The Dow Jones Industrial Average plummeted by 800 points， the S&P 500 fell 1.04% in a single day， and the Software ETF dropped over 4%.”
Leading the decline were software stocks such as Microsoft， Oracle， and Accenture， as well as payment giants like Visa， Mastercard， and American Express， with drops ranging from 2% to 7%. Even DoorDash experienced a significant decline due to predictions of potential disruption by AI delivery agents.
However， Citrini’s report was quickly met with strong opposition from numerous economists. The Chair of the White House Council of Economic Advisers even described it as “an interesting piece of science fiction” during a press conference. Citadel Securities also published an article refuting the report’s claims， arguing that AI is more likely to complement rather than replace the workforce and that its diffusion rate is limited by physical and economic laws， preventing exponential growth.
What Exactly Did Citrini’s Report Claim?
At the outset， the author preemptively stated that the report presented a hypothetical scenario rather than a firm prediction. This qualification was likely due to the alarmist nature of the depicted future.
Citrini posited that the US economy is fundamentally a white-collar service-based economy， with white-collar workers constituting 50% of total employment and contributing approximately 75% of disposable consumer spending. The report predicted that as AI technology matures， companies would lay off large numbers of high-paid white-collar employees， replacing them with cheaper AI solutions， thereby triggering a chain reaction.
The report theorized that reduced consumer spending by laid-off white-collar workers would lead to decreased corporate revenues. This， in turn， would further incentivize companies to adopt more AI， resulting in more job losses and a vicious cycle. The displaced white-collar workers would then be forced into lower-paying service and gig economy jobs. For instance， a former senior product manager at Salesforce， earning $180，000 annually， might end up driving for Uber with an income as low as $45，000.
The prospect paints a grim picture， indeed.
Citrini also coined the term “Ghost GDP” to describe a situation where corporate profits increase， but individuals do not benefit， leading to reduced purchasing power and slower circulation of money. The report further predicted that the collapse of the real economy would spread to the financial sector， with the S&P 500 index declining by nearly 40% from its peak by 2028. With white-collar workers unable to make loan payments， the report projected widespread defaults on prime mortgages， leading to a housing market collapse in tech hubs like San Francisco and Seattle. The report concluded with a dramatic prediction: by June 2028， US unemployment would surge to 10.2%， and angry unemployed individuals would launch an “Occupy Silicon Valley” movement， blockading the major AI labs.
What Were the Flaws in the Report?
Citrini’s report， while causing market jitters， faced significant criticism from economists who argued it “completely disregarded economic principles.” Citadel Securities’ research highlighted that AI development is constrained by physical and economic laws， meaning its adoption rate would not be exponential and therefore not as catastrophic as predicted.
Firstly， Citrini’s assumption that white-collar professionals like software engineers would be replaced by AI was challenged. Citadel pointed out that demand for software engineers has actually been increasing， with job openings growing by 11% year-over-year. This aligns with historical patterns; for example， the prediction that AI would render radiologists obsolete a decade ago has not materialized; instead， radiologists remain in high demand.
Furthermore， prominent industry figures like Jensen Huang， Demis Hassabis， and Linus Torvalds have repeatedly emphasized that roles like software engineers and programmers are unlikely to disappear. AI is viewed as a tool， and the future will require individuals who can effectively leverage it.
Secondly， the report conflated recursive technological capability with recursive adoption. While AI’s self-improvement capabilities are impressive and its development rapid， the pace of economic deployment is considerably slower. Historical evidence suggests that technological diffusion follows an S-curve: slow in the initial phase， accelerating in the middle， and eventually plateauing due to cost and regulatory saturation. This implies that while AI is advancing rapidly， it will eventually encounter limitations such as insufficient energy， prohibitive computing costs， or stricter regulations， allowing humanity time to adapt.
Moreover， AI operation relies on fundamental resources: chips (semiconductors)， data centers， and electricity. In a scenario where numerous companies attempt to replace human workers with AI， there would be an intense competition for these resources， driving up their prices. If the cost of using AI surpasses that of human labor， companies would logically opt to hire people， thus preventing widespread AI-driven displacement.
Citadel concludes that AI is more likely to act as a complement to the workforce rather than a substitute. Similar to how Microsoft Office reshaped tasks and increased efficiency without eliminating office workers， AI is expected to augment human capabilities and improve productivity rather than eradicate jobs. While AI will undoubtedly create a productivity shock， it will be a positive one， lowering marginal costs， expanding potential output， and increasing real incomes. This pattern has been observed throughout history with every major technological advancement， from steam power and electrification to computers.
Conversely， AI is also poised to spur job creation. For instance， the expansion of data centers has led to increased hiring in the construction sector.
One More Thing
As the situation has unfolded， market fear has subsided， leading to a collective sigh of relief. The realization that the initial panic was potentially unfounded has led to a rebound in stocks that were unfairly impacted.
On the stock market， companies that experienced sell-offs on Monday saw a strong rebound on Wednesday and Thursday. Microsoft’s stock price quickly recovered to around $389， and Oracle’s stock surged by nearly 10%. Payment giants Visa and Mastercard also regained most of their losses， bolstered by unexpectedly strong consumer confidence data released on Tuesday (91.2， exceeding the forecast of 87.1).
Interestingly， at the time of Citrini’s report release， founder James Van Gellen expressed shock at the stock sell-off and admitted he “hoped he was wrong” and was “glad someone tried to debunk it.” He humorously added that if he had known the report would influence the stock market， he would not have published it for free.
Since the beginning of the year， with Anthropic’s industry plugins causing software stocks to falter， market sentiment has been on edge. The high valuations of tech stocks and crowded positions may have made the market susceptible to herd behavior triggered by even minor news.
However， as Jensen Huang has repeatedly stated， the market has fundamentally misunderstood the threat AI poses to software companies. AI is not expected to replace existing software tools; instead， it will become an integral part of their usage， significantly enhancing development and operational efficiency for software firms.
During the dot-com bubble burst at the end of the 20th century， similar predictions of the internet being a “Ponzi scheme” were made. History has shown that after the bubble， the underlying infrastructure remained， and it went on to reshape the world.
Reports like Citrini’s， with their doomsday scenarios， should be treated as “science fiction.”
