Andrew Leigh warns AI may lift wages yet widen wealth gap

Leigh’s message lands at a moment when frontier AI capability is racing ahead of labour market policy. His core point is subtle but important: AI can simultaneously compress gaps in individual productivity while amplifying the returns to owning capital in AI firms and infrastructure. That breaks the simple story that “AI either helps or hurts inequality” and instead forces policymakers to separate the distribution of skills from the distribution of assets.

For the race to AGI, this framing matters because it anticipates a world where very powerful models are widely deployed long before fully automated workers exist. In that middle zone, highly capable models can boost output for average employees, but the main financial upside accrues to whoever owns the models, chips and platforms. Unless tax, competition and ownership rules adjust, the result is a thinner middle class and even more concentrated control over future AI development.

Leigh is also implicitly pushing against techno determinism. He argues that whether AI deepens or narrows inequality depends on competition policy, worker bargaining power and how governments choose to tax and re invest AI driven profits. That is a reminder to AGI watchers that distributional outcomes are not fixed by model weights; they are shaped by political choices that can lag years behind the technology.