Anthropic Maps Three Ways AI Could Reshape the Economy

Anthropic has published an interactive model of how artificial intelligence might reshape the American economy by 2030, and the most useful thing about it is that it refuses to give a single answer. Released on 9 September by the company's economics group alongside a technical report titled Economic Scenarios for Transformative AI, the tool lets anyone move the assumptions and watch the numbers shift across three broad futures.

Three futures, one dashboard

In the modest scenario, AI turns out to matter about as much as the internet did. US gross domestic product reaches roughly $34.1 trillion by 2030, around 1.6 percent above the baseline projection, and the labour market absorbs the change without much drama.

The substantial scenario is where the picture gets more pointed. Here AI handles close to half of all knowledge work by 2030. Output climbs to about $36.3 trillion and growth roughly doubles its normal rate, yet the gains do not land evenly. Unemployment settles near 5 percent while wages for knowledge workers stay flat, the economy expanding around the people whose jobs it is quietly absorbing.

The extreme scenario pushes both dials hard. GDP runs 32.4 percent above baseline, near $44.4 trillion, a figure that would represent one of the fastest expansions in modern history. In the same run, pay for knowledge workers falls by more than 10 percent, and some versions of the model show unemployment climbing well past postwar norms. Prosperity and displacement arrive together, which is the uncomfortable point the whole exercise is built to make.

What ordinary people expect

Anthropic paired the modelling with a survey of 10,980 US adults, conducted through Morning Consult between 11 and 23 August. The typical respondent landed somewhere in the middle, implying GDP about 10 percent higher by 2030 than it would be without AI, with unemployment around 5 percent. Roughly one in ten held views close to the extreme scenario. The report, summarised by Unite.AI, frames these as scenarios to reason about rather than forecasts to bet on.

Reading it honestly

A model like this is only as good as its assumptions, and Anthropic is transparent that it is offering a structured way to argue rather than a prediction. Still, it sits usefully against the harder data we already have. Gartner recently found that only 22 percent of organisations have actually scaled AI across their operations, a reminder that the substantial and extreme paths both assume a pace of adoption the real economy has not yet shown. The questions of whether AI is already gutting the workforce or is about to replace programmers remain open. What Anthropic's explorer does well is force a reader to hold two ideas at once, that AI could make the country meaningfully richer and meaningfully more unequal, and that which of those dominates is not yet decided.