CEOs Report Minimal AI Impact on Productivity and Employment in Recent Study

Recent research shows most corporate executives aren't seeing measurable productivity gains from AI implementation despite widespread adoption talk. A National Bureau of Economic Research study published in February surveyed 6,000 CEOs, CFOs, and other executives from firms in the U.S., U.K., Germany, and Australia.
Key Findings from the Data
- About two-thirds of executives reported using AI, but usage averaged only 1.5 hours per week
- 25% of respondents reported not using AI in the workplace at all
- Nearly 90% of firms said AI has had no impact on employment or productivity over the last three years
- 374 S&P 500 companies mentioned AI in earnings calls, with most reporting positive implementation, but this isn't reflected in broader productivity gains
The Solow Productivity Paradox Comparison
Economists are noting similarities to Robert Solow's 1987 observation about computers: "You can see the computer age everywhere but in the productivity statistics." Following the advent of transistors, microprocessors, integrated circuits, and memory chips in the 1960s, productivity growth actually slowed from 2.9% (1948-1973) to 1.1% after 1973.
Apollo chief economist Torsten Slok wrote: "AI is everywhere except in the incoming macroeconomic data. Today, you don't see AI in the employment data, productivity data, or inflation data." He added that outside of the Magnificent Seven tech companies, there are "no signs of AI in profit margins or earnings expectations."
Conflicting Research and Future Expectations
- The Federal Reserve Bank of St. Louis observed a 1.9% increase in excess cumulative productivity growth since ChatGPT's late-2022 introduction
- A 2024 MIT study found a more modest 0.5% increase in productivity over the next decade
- MIT researchers claimed in 2023 that AI could increase worker performance by nearly 40% compared to non-AI users
- Executives forecast AI will increase productivity by 1.4% and output by 0.8% over the next three years
- Firms expect a 0.7% cut to employment over this period, while individual employees surveyed saw a 0.5% increase
ManpowerGroup's 2026 Global Talent Barometer found that across nearly 14,000 workers in 19 countries, regular AI use increased 13% in 2025, but confidence in the technology's utility plummeted 18%, indicating persistent distrust.
📖 Read the full source: HN AI Agents
👀 See Also

Stanford Report Shows AI Experts and Public Have Diverging Views on AI Impact
Stanford's annual AI industry report reveals significant gaps between AI experts' optimism and public anxiety, with experts focusing on AGI risks while the public worries about jobs, medical care, and utility costs.

Visual Reasoning Benchmark Results for 15 Multimodal AI Models
AIMultiple benchmarked 15 leading multimodal AI models on 200 visual reasoning questions across two tracks: chart understanding and visual logic. Gemini-3.1-pro-preview and Gemini-3-pro-preview lead the overall results, followed by GPT-5.2, Kimi-K2.5, and GPT-5.2-pro.

Claude Opus 4.1 scores 17.75% on SWE-Bench Pro's private dataset, highlighting memorization vs. reasoning gap
Claude Opus 4.1 scored 80% on SWE-Bench Verified but dropped to 17.75% on SWE-Bench Pro's private dataset of 276 tasks from 18 proprietary startup codebases. Scale AI's analysis found models were navigating by memory rather than reasoning on familiar repositories.

Claude Code: Feedback Honeypot Overrides Privacy Opt-Out — Users Report Session Transcript Trap
Anthropic's Claude Code now prompts users to allow session transcript review — pressing 'n' for no logs 'Thanks for your feedback' and may still train models. Dismiss key behavior is unclear.