Anthropic, a leading AI research organization, has built a model to predict how AI might impact the US economy by 2030. The model, released on September 10, presents three possible scenarios: a modest boost, substantial growth, and an extreme case with significant job loss. According to the model, the most extreme scenario could result in an economy roughly a third larger than it would be without AI, but with knowledge-work unemployment nearing 18 percent.
The model treats every job as a bundle of tasks, with AI potentially leaving a task alone, augmenting it, automating it, or adding new ones. The effects are then scaled across the economy under different assumptions about AI's capabilities and adoption rates. The result is three distinct AI economic scenarios, labeled modest, substantial, and extreme.
In the modest scenario, AI has a similar impact to the internet, with US GDP reaching $34.1 trillion by 2030, about 1.6 percent above a no-AI path. The substantial scenario shows AI doing half of all knowledge work by 2030, mostly autonomously, resulting in an economy growing at twice its normal rate to $36.3 trillion. Wages for knowledge workers remain flat, while other workers see gains.
The extreme scenario is the most striking, with AI being more productive than humans at most knowledge-work tasks and doing nearly all of them autonomously. This path would likely require recursively self-improving systems, adopted quickly, and could lead to annual GDP growth of 15 percent, doubling the economy every four and a half years. However, this scenario also predicts significant job loss, with unemployment among knowledge workers rising to about 18 percent by 2030.
Anthropic's model notes that switching occupations is slow and often leaves people out of work for a long time. The company surveyed over 10,000 Americans in August, with the typical respondent's answers implying a scenario close to the substantial case. About 10 percent of respondents gave answers in line with the extreme scenario.
Economists have pushed back against the model's predictions, arguing that double-digit growth rates are unlikely. Ben Moll and Alex Imas have placed a public bet that US per-capita real GDP growth will stay below 15 percent every year through 2033, citing the physical nature of most economic work and the time it takes to build machines and data centers.
Anthropic's CEO, Dario Amodei, has said AI could push growth to 10 to 15 percent a year, but the company's economists acknowledge that the model is a simplification, leaving out policy responses, business cycles, and aggregate demand effects. The model's authors say they are not forecasting and will continue to revise the model as evidence develops.
This article was written with the assistance of AI.
News Factory APP - agentic news to boost your SEO & AEO.