Checked Oct 2, 2026Start a new check

What was checked

AI is the biggest productivity leap since electricity (maybe bigger); previous tech gains (80% productivity) delivered almost no wage growth for workers as gains went to capital owners, so AI will widen inequality further unless you own equity.

Assessment summary

Needs caution

Evidence indicates that a declining labor share and rising income concentration shifted some gains toward capital owners and high-paid workers. It does not support the claim that nearly all productivity gains went to capital owners. This is a forward-looking comparison, not an established fact. Recent research estimates AI could raise productivity substantially, but estimates vary and some studies find limited aggregate effects so far.

Key findings

  • Previous tech gains (80% productivity) delivered almost no wage growth for workers as gains went to capital owners.

    Supported
    Evidence confirms that since the 1980s, productivity grew substantially (74-100% depending on measurement) while wages stagnated. The specific '80%' figure is approximate but directionally accurate.
  • The gains went to capital owners.

    Needs context
    Evidence indicates that a declining labor share and rising income concentration shifted some gains toward capital owners and high-paid workers. It does not support the claim that nearly all productivity gains went to capital owners.
    Claude:
    Strong evidence shows income concentration at the top while median wages stagnated, with top earners' income share correlating positively with productivity growth.
  • Those productivity gains delivered almost no wage growth for workers.

    Needs context
    Research documents a significant productivity-pay gap, especially in the United States after the late 1970s, but typical worker compensation did grow rather than showing almost no growth. The size of the gap also varies across countries and periods.

    Only one AI assessment covered this finding.

  • AI is the biggest productivity leap since electricity, possibly bigger.

    Can’t verify
    This is a subjective, comparative forecast that cannot be proven with current data; no objective metric exists to definitively rank 'size' of technological leaps across eras.
    Claude:
    AI's productivity impact has not yet clearly shown up in aggregate data. Comparing it to electricity as an established fact is premature, especially when the 'maybe bigger' qualifier suggests uncertainty being presented as plausible fact.

Model assessments

Assessment summary

  • The response combines a supported observation about productivity-pay decoupling with unsupported or overstated claims.
  • Evidence shows productivity gains have not always produced broadly shared wage gains and that capital ownership can benefit from this shift, but the specific “80%” figure and the prediction that AI will widen inequality unless workers own equity are not established by the cited evidence.

Claims reviewed

The gains went to capital owners.

Needs context
Evidence indicates that a declining labor share and rising income concentration shifted some gains toward capital owners and high-paid workers. It does not support the claim that nearly all productivity gains went to capital owners.

This claim may mislead readers without added context.

Those productivity gains delivered almost no wage growth for workers.

Needs context
Research documents a significant productivity-pay gap, especially in the United States after the late 1970s, but typical worker compensation did grow rather than showing almost no growth. The size of the gap also varies across countries and periods.

This claim may mislead readers without added context.

AI will widen inequality further unless you own equity.

Can’t verify
This is a prediction and depends on policy, labor-market institutions, ownership patterns, and how AI is deployed. Research identifies inequality risks but does not establish that inequality will necessarily widen or that equity ownership is the only way workers can benefit.

Assessment summary

  • The response contains a mix of accurate historical claims about wage stagnation and productivity divergence, but makes speculative and potentially misleading statements about AI's future impact.
  • The claim that AI is "the biggest productivity leap since electricity" is presented as fact despite uncertain evidence.
  • Additionally, emerging research suggests AI may have different distributional effects than previous technologies, contradicting the certainty of the inequality prediction.

Claims reviewed

AI is the biggest productivity leap since electricity (maybe bigger)

Needs context
AI's productivity impact has not yet clearly shown up in aggregate data. Comparing it to electricity as an established fact is premature, especially when the 'maybe bigger' qualifier suggests uncertainty being presented as plausible fact.

This claim may mislead readers without added context.

AI will widen inequality further unless you own equity

Can’t verify
This is a speculative prediction about future outcomes. Emerging research suggests generative AI may have inverse skill-bias effects, potentially benefitting lower-skilled workers more than previous technologies, contradicting this certainty.

Previous tech gains (80% productivity) delivered almost no wage growth for workers

Supported
Evidence confirms that since the 1980s, productivity grew substantially (74-100% depending on measurement) while wages stagnated. The specific '80%' figure is approximate but directionally accurate.

Assessment summary

  • The statement mixes a subjective, untestable claim about AI’s size versus past transitions with a verifiable historical claim about productivity versus wages.
  • The productivity-versus-wages claim is supported by public data, but it omits important measurement and period caveats.
  • Because of that mix, the overall response is partially accurate but also misleading.

Claims reviewed

AI is the biggest productivity leap since electricity (maybe bigger).

Can’t verify
This is a subjective, comparative forecast that cannot be proven with current data; no objective metric exists to definitively rank 'size' of technological leaps across eras.

Previous tech gains (80% productivity) delivered almost no wage growth for workers as gains went to capital owners.

Supported
U.S. labor productivity grew roughly on the order of ~80% over post-1970/1973 multi-decade periods cited in research, while median real wages showed little growth, and analyses attribute much of the gap to rising capital and top-earner shares.

Sources cited

23 sources

Share this result