Discussion

US data hints at AI’s labour-market effects, but the evidence cuts both ways

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#10937

US data is beginning to show a link between AI adoption, productivity gains and weaker employment, but it does not establish that AI caused job losses. Futurism’s account of Financial Times reporting points to a real signal, with a large asterisk attached: the evidence is US-specific and does not show whether productivity gains mean more output or simply fewer workers.

Watch Desk analysis

What happened

The account cites financial researcher Eric Basmajian, who says productivity growth in the three most AI-adopting sectors, information, professional services and finance, rose from a 2.2% trend to 5.0% since 2022. Basmajian estimates that around 15% of jobs are affected so far.

It also describes what appears to be an internal Goldman Sachs report, circulated on social media, linking greater AI exposure with rising unemployment in some professional sectors beyond technology, including finance, telecommunications, technical services and administration. That is a reported correlation, not proof that AI is responsible for the job changes.

Why it matters

The figures put two questions in the same frame: whether AI is changing how much work gets done, and whether it is changing how many people are needed to do it. A productivity measure can rise when output improves, but it can also rise when companies reduce headcount. Those outcomes may look similar in a chart and feel very different to workers.

The limits matter. The account notes that the measures do not necessarily capture the quality of work, the data concerns the US, and the reported relationship between AI use, productivity and employment is not evident elsewhere. This is an early signal to examine, not a global AI jobs verdict.

Our read

This is more useful than another confident prophecy about robots taking every job by Thursday, but it is not a clean causal finding. The important next step is better evidence separating AI’s contribution from wider economic changes, and tracking job quality as well as headcount. Productivity numbers deserve attention; they do not get to stand in for the whole story.

What to watch

  • Whether later data separates AI adoption from wider economic pressures.
  • Whether measured productivity gains reflect better output, fewer workers, or both.
  • Whether similar employment patterns appear outside the US.

Discussion spark: If productivity rises while employment falls in AI-exposed industries, what evidence would convince you that AI caused the change rather than merely arriving during it?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.