Thousands of CEOs just admitted AI had no impact on employment or productivity
Source: Hacker News
The Productivity Paradox Re‑emerges with AI
The original paradox (1987)
- Robert Solow, Nobel‑winning economist, noted that after the rapid advances of the 1960s (transistors, microprocessors, integrated circuits, memory chips), productivity growth slowed dramatically.
- From 2.9 % (1948‑1973) to 1.1 % (post‑1973)【source】(https://www.brookings.edu/wp-content/uploads/2016/06/199904.pdf).
- Instead of boosting output, early computers often generated excess information, producing detailed reports that filled reams of paper.
- Solow summed it up:
“You can see the computer age everywhere but in the productivity statistics.”
— New York Times Book Review, 1987【source】(https://www.standupeconomist.com/pdf/misc/solow-computer-productivity.pdf)
AI today: a similar story?
- 374 S&P 500 firms mentioned AI in earnings calls (Sept 2024‑2025). Most described the technology’s implementation as entirely positive, yet broader productivity gains remain elusive【source】(https://www.ft.com/content/e93e56df-dd9b-40c1-b77a-dba1ca01e473).
- A National Bureau of Economic Research (NBER) working paper (April 2024) surveyed ~6,000 CEOs, CFOs, and other executives across the U.S., U.K., Germany, and Australia【source】(https://www.nber.org/papers/w34836). Key findings:
| Metric | Result |
|---|---|
| Executives reporting AI use | ≈ 66 % |
| Average AI usage per week | 1.5 hours |
| Executives reporting no AI use | 25 % |
| Firms seeing no impact on employment or productivity (last 3 years) | ≈ 90 % |
| Forecasted AI‑driven productivity increase (next 3 years) | +1.4 % |
| Forecasted AI‑driven output increase (next 3 years) | +0.8 % |
| Expected employment change (firm‑level) | ‑0.7 % |
| Expected employment change (individual‑level) | +0.5 % |
What the data suggest
- Limited current impact – Most executives are only dabbling in AI (≈ 1.5 h/week), and a quarter still do not use it at all.
- High expectations for the future – Despite modest present‑day usage, firms anticipate productivity gains of around 1–1.5 % over the next three years.
- A repeat of Solow’s paradox? – As with early computers, the observable productivity statistics lag behind the hype surrounding a transformative technology.
Take‑away
- History may be repeating itself: New technology (computers → AI) is highly visible and widely praised, yet aggregate productivity metrics have yet to reflect the promised boost.
- Policymakers and managers should temper expectations, focus on effective integration, and monitor real‑world output rather than relying solely on anecdotal success stories.
Solow Strikes Back
In 2023, MIT researchers claimed AI implementation could increase a worker’s performance by nearly 40 % compared with workers who didn’t use the technology — see the MIT Sloan study. Yet emerging data that fail to show these promised productivity gains have left economists wondering when—or if—AI will deliver a return on corporate investments, which swelled to more than $250 billion in 2024 (Stanford AI Index).
“AI is everywhere except in the incoming macro‑economic data,” wrote Apollo chief economist Torsten Slok in a recent blog post, invoking Solow’s observation from nearly 40 years ago. “Today, you don’t see AI in the employment data, productivity data, or inflation data.”
Slok added that outside of the Magnificent Seven, there are no signs of AI in profit margins (source) or earnings expectations (source).
Academic Findings: A Contradictory Picture
- Federal Reserve Bank of St. Louis – In its State of Generative AI Adoption report, the bank observed a 1.9 % increase in excess cumulative productivity growth since the late‑2022 introduction of ChatGPT (link).
- MIT (2024) – A study found a more modest 0.5 % productivity gain over the next decade (link).
“I don’t think we should belittle 0.5 % in 10 years. That’s better than zero,” said study author and Nobel laureate Daron Acemoglu. “But it’s just disappointing relative to the promises that people in the industry and in tech journalism are making.”
Emerging Research on Workforce Attitudes
- ManpowerGroup’s 2026 Global Talent Barometer surveyed nearly 14 000 workers in 19 countries. It found that while regular AI use rose 13 % in 2025, confidence in the technology’s utility plummeted 18 %, indicating persistent distrust (Fortune coverage).
- IBM – Chief Human Resources Officer Nickle LaMoreaux announced that IBM will triple its number of young hires. The move suggests that, despite AI’s ability to automate many entry‑level tasks, displacing those workers could create a shortage of middle managers and jeopardize the company’s leadership pipeline (Fortune article).
The Future of AI Productivity
The productivity pattern we see today could reverse. The IT boom of the 1970s‑and‑80s eventually gave way to a surge of productivity in the 1990s and early 2000s, including a 1.5 % increase in productivity growth from 1995 to 2005 after decades of slump【Brookings】.
Key Observations
| Analyst / Source | Main Point | Evidence |
|---|---|---|
| Erik Brynjolfsson – Economist, Stanford Digital Economy Lab | The current productivity surge may already be reversing. | Fourth‑quarter GDP was up 3.7 %, even though the jobs report revised weekly gains down to 181,000. His own analysis shows a 2.7 % U.S. productivity jump last year, driven by the transition from AI investment to benefit realization【Financial Times Op‑Ed】. |
| Mohamed El‑Erian – Former Pimco CEO & Economist | Job growth and GDP growth are continuing to decouple, partly due to AI adoption—mirroring the 1990s office‑automation wave【FT】. | |
| Torsten Slok – Industry commentator | AI’s impact may follow a “J‑curve”: an initial slowdown followed by an exponential surge. The shape depends on the value created by AI. | |
| Torsten Slok (cont.) | Unlike the 1980s IT market—where innovators enjoyed monopoly pricing until rivals caught up—today AI tools are widely accessible because fierce competition among large‑language‑model builders drives prices down. | |
| Torsten Slok (cont.) | The future of AI productivity hinges on how companies adopt and integrate generative AI, not merely on the product itself. |
The “J‑Curve” Narrative
“From a macro perspective, the value creation is not the product, but how generative AI is used and implemented in different sectors of the economy.” – Torsten Slok
- Initial slowdown: Early adoption may produce modest or even negative returns as organizations experiment and adjust processes.
- Exponential surge: Once best‑practice use cases emerge and integration deepens, productivity can accelerate dramatically.
What Determines the Trajectory?
- Competitive Landscape – Intense rivalry among LLM providers is lowering costs and expanding access.
- Corporate Adoption – Firms that actively embed AI into workflows (e.g., R&D, customer service, supply‑chain planning) will capture the bulk of productivity gains.
- Sector‑Specific Implementation – Value creation varies widely across industries; the same AI tool can have negligible impact in one sector and transformative effects in another.
Takeaway
The next wave of AI‑driven productivity will likely mirror a J‑curve, with the magnitude of the surge depending less on the technology itself and more on strategic, sector‑wide implementation. Companies that prioritize integration and innovation stand to reap the greatest benefits.