Could AI's automation of clerical work end up expanding bureaucratic tasks and cognitive load instead of cutting them? The so-called "Jevons Paradox" has emerged as a new focal point in the productivity debate of the generative-AI era.
Feb 26, 2026 · AI & The Workplace
The Great Productivity Panic of 2026
Agentic coding tools like Claude Code promised to lighten the load. A year into "vibe coding," the verdict from the ground is blunter: AI doesn't reduce work — it intensifies it.
~80%
Estimated time cut on a typical 90-min task (Anthropic)
3×
Median speed gain reported by 349 technologists (METR 2026)
77%
Workers who said AI actually lowered their productivity (Upwork)
The numbers don't agree
Annual labor-productivity boost from AI — estimates by source. Same metric, very different stories.
ARK / Cathie Wood"already"
Anthropic10-yr est.
Goldman / OECDhigh end
Goldman / OECDlow end
The intensification loop
Why faster tools haven't meant lighter days — the cycle a Berkeley study traces.
AI accelerates each task
→
Expectations rise
→
Scope & hours expand
→
Burnout
Speed up
Great for boilerplate, tests and scaffolding — developers report shipping 2–3× faster .
53% saw improved employee productivity · 33% hourly gain in one study
Slow down
Speed rises but comprehension drops. An Anthropic RCT (52 learners) found outcomes 17% lower .
Another analysis: overall time saved just 1.4%
The open question isn't the technology's raw power — it's whether organizations adapt to it. 79% of Claude Code's use leans toward automation, but workflows and expectations haven't caught up.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…