WEF vs. Anthropic: Comparing Major New Reports on AI & the Future of Work
- Chris McLellan
- 23 hours ago
- 5 min read
Ask AI dives into new reports from WEF and Anthropic, both published in January 2026, to understand where they align, where they disagree, and why it matters now
Quick Summary
Different approaches: The WEF report is based largely on executive surveys and future scenarios, while Anthropic focuses on real-world Claude usage data from people and businesses.
Where they agree: AI benefits are uneven and currently favor highly skilled workers and wealthier regions. Both reports link AI use to education and income levels (Anthropic, p. 6; WEF, p. 168).
Where they diverge: The WEF finds 54% of executives expect AI to displace jobs, and only 12% expect it to increase wages, while Anthropic’s usage data shows most AI today is used to support people rather than replace them (WEF, p. 164; Anthropic, p. 107).
Impact on productivty: Anthropic estimates AI could lift annual productivity by 1.8%, but adjusting for errors cuts that to about 1.0%, highlighting reliability as a key constraint (Anthropic, p. 107).
The quick math: A 1% increase in U.S. productivity would translate into roughly $250–300 USD billion in additional economic output each year. (Ask AI estimate)
Breaking news: Anthropic engineers have claimed that Claude built their massively-hyped Cowork platform in just over a week, mostly (if not entirely) on its own.
Download these reports:
Where Anthropic & WEF Reports Align
Despite their different approaches, Anthropic and the World Economic Forum land on a shared conclusion:
AI adoption is uneven, and it is currently reinforcing existing advantages rather than closing gaps.
Anthropic’s Economic Index shows that AI usage closely tracks income levels. While adoption across U.S. states may even out within two to five years, global use remains heavily concentrated in richer countries (Anthropic, p. 6). In other words, access to AI today largely mirrors where economic power already sits.
The WEF echoes this concern from a workforce and policy perspective. It warns that uneven adoption can “fuel inequality, create a bifurcated economy and limit growth” if skills development and readiness fail to keep pace with technology (WEF, p. 168).
Education is one of the clearest fault lines.
Anthropic finds that the tasks people most often use AI for tend to require more education than the average task in the economy (Anthropic, p. 10).
The WEF sees the same pattern playing out in hiring demand, reporting a 70% increase in demand for AI literacy skills between 2024 and 2025 (WEF, p. 175).
Taken together, both reports suggest the early gains from AI are accruing to people, roles, and regions that already have stronger skills and resources, at least for now.
Where executives diverge from AI useage data
Both reports agree AI is changing work, but they tell very different stories about the scale, speed, and shape of that change.
WEF Report: AI & The Future of Work
Drawing on executive surveys and future scenarios, the World Economic Forum report highlights the risk of disruption:
54% of the executives polled expect AI to displace existing jobs (WEF, p. 164).
Only 12% expect AI to lead to higher wages (WEF, p. 164).
Outcomes depend heavily on how well skills development, governance, and workforce transitions are managed, with futures ranging from rapid progress to stalled adoption (WEF, pp. 166–167).
What Anthropic's Report Highlights
Looking at how people and businesses actually use AI today, Anthropic paints a more constrained picture:
Most AI use currently supports people rather than replacing them outright.
AI saves time, but reliability drops as tasks become longer or more complex, limiting full automation.
When errors and rework are accounted for, estimated gains in U.S. labor productivity growth fall from 1.8% to about 1.0% (Anthropic, p. 107).
In some jobs, AI takes over the more complex parts of the work, which can lower the overall skill level of the role. In others, it removes the boring, repetitive tasks and leaves people focused on decisions and judgment. Travel agents and property managers show how these paths can differ (Anthropic, pp. 11–12).
Anthropic emphasizes that this is not a forecast of job losses, but a snapshot based on which tasks AI supports today (Anthropic, Ch. 4).
Breaking News: Claude Codes Platform In 12 Days
Around the same time these reports were released, Anthropic shared a concrete example of how quickly AI-assisted work can now move. The company revealed that much of its newly launched Cowork tool was built using Claude itself.
When asked on X how much of the new Claude Cowork tool was built using Anthropic's AI coding agent (Claude Code), Boris Cherny, Anthropic's head of Claude Code, replied: "All of it".
What matters here isn’t speed, but how the work changes. If AI can build a major digital product in days, human effort shifts away from execution and toward direction, review, and judgment. Both reports point to the same pattern.
Why This Matters
Taken together, these reports show that AI is not a rising tide. It is creating clear winners and losers, at least in the near term. People with stronger skills, education, and access are seeing the benefits first, while others face job loss, shrinking roles, or fewer paths to advancement.
Education increasingly looks like the main lifeboat. Both reports point to skills and learning as the deciding factor in whether AI amplifies opportunity or deepens inequality. That also makes broad, one-size-fits-all policies hard to design, since the effects of AI vary widely by role, sector, and region.
There is also a quieter risk, especially for younger workers. As AI takes on more of the “doing,” there may be a temptation to skip foundational skills that are essential for reviewing, correcting, and supervising AI outputs. Short-term productivity gains can mask longer-term damage to skill development.
Ask AI has asked before, and asks again:
Is productivity the correct primary measure of AI success, or should we give equal weight to social outcomes like skill development, trust, and the long-term health of the workforce, especially for younger generations?
About Ask AI
Since 2017, Ask AI has operated as an independent nonprofit helping people spot the opportunities and navigate the challenges associated with the increased adoption of artificial intelligence at work.
Our mission is to be an idependent voice sharing diverse perspectives on this transformative technology that is reshaping the future of work and redefining professions and industries worldwide. Our volunteers produce a interviews, short videos and free guides.
Looking for AI industry experience? Consider volunteering
Curious about AI and future of work? Join newsletter
Event organizer? Let's collaborate