Despite widespread anticipation of a transformative economic impact, artificial intelligence has yet to demonstrably boost macroeconomic indicators, according to Apollo Global Chief Economist Torsten Slok. The observation echoes a sentiment expressed decades ago as the personal computer revolution unfolded, when economist Robert Solow famously noted that the benefits of the new technology were “everywhere but in the productivity statistics.”
In a note published on , Slok highlighted the continued absence of AI’s influence in key economic data points – employment figures, productivity growth, and inflation rates. He further noted that profit margins and earnings forecasts for S&P 500 companies, excluding the so-called “Magnificent 7” technology giants, also show no clear evidence of an AI-driven uplift.
“AI is everywhere except in the incoming macroeconomic data,” Slok wrote, encapsulating the current disconnect between the hype surrounding the technology and its measurable economic effects.
The lack of broad economic impact hasn’t dampened investor enthusiasm, however. Recent market volatility has seen significant sell-offs in companies perceived as vulnerable to disruption by AI, including wealth managers, insurance brokerages, tax preparation firms, accounting services, data analytics companies, legal research providers, and logistics and trucking businesses.
This investor apprehension contrasts sharply with the optimistic projections offered by AI proponents. Anthropic CEO Dario Amodei, speaking at the World Economic Forum last month, predicted that AI could drive annual GDP growth by as much as 5% to 10%. Elon Musk, cofounder of xAI, went even further, suggesting that AI will ultimately create so much wealth that work itself could become optional within the next two decades.
Slok remains skeptical of these more exuberant forecasts. He acknowledges the possibility of a “J-curve effect,” where the economic benefits of AI take time to materialize and become visible in macroeconomic data. However, he cautions that this outcome is not guaranteed.
A key difference between the current AI revolution and the computer revolution of the 1980s, Slok points out, lies in the dynamics of value creation. Unlike the early days of computing, where innovators enjoyed monopoly pricing power, the current landscape of large language model development is characterized by intense competition, driving prices for end-users toward zero.
From a macroeconomic perspective, the value of AI is determined by its application within the broader economy, not by the technology itself. Studies to date offer a relatively modest outlook. The Penn Wharton Budget Model, for example, forecasts an annual increase in total factor productivity of just 0.1 to 0.2 percentage points, translating to a cumulative boost of only 1.5% by .
The Congressional Budget Office (CBO) offers a similarly conservative assessment, estimating that AI will add just 0.1 percentage point per year to total factor productivity growth, eventually boosting output by 1 percentage point by . This projection comes as the Labor Department revised its initial estimate of job gains in down to 181,000, a significant reduction from the initial print of 584,000 and the 1.46 million jobs added in .
The unexpectedly low job growth in the face of continued economic expansion raises questions about the impact of AI on productivity. The CBO suggests that the widespread adoption of generative AI applications is “expected to improve business efficiency and the organization of work and thus to lift TFP growth modestly over the next decade.”
Slok’s analysis suggests that AI is more likely to be “labor enhancing” in certain sectors rather than “labor replacing” across the board. This nuanced view aligns with observations that AI is currently generating “micro-productivity” gains – small, incremental improvements in efficiency across a wide range of tasks – rather than triggering a large-scale overhaul of the economy. As one LinkedIn commentator noted, the current state of AI is akin to the “Green Screen Era” of computing, where the benefits were real but not immediately reflected in national productivity statistics.
The debate over AI’s economic impact underscores the challenges of measuring the benefits of transformative technologies. As with previous technological revolutions, the full effects of AI may take years, or even decades, to fully materialize and become apparent in macroeconomic data. For now, the evidence suggests that the much-hyped AI revolution is still waiting for its moment to arrive.
