Anthropic Models Three Scenarios for the U.S. Economy by 2030

• Anthropic has forecasted the U.S. economy through 2030.
• Under one of the three scenarios—the most radical—GDP is expected to grow by 15% per year, but unemployment is also anticipated to reach record levels.
• In turn, workers may lose part of their income, while a significant portion of profits could shift to capital owners.

Anthropic has published a model of potential economic consequences of AI development by 2030. The company considered three scenarios—from a moderate impact similar to the internet effect to a radical transformation of the economy where AI would perform nearly all intellectual work.

The most extreme scenario predicts U.S. GDP growth of up to 15% per year, but at the same time, a sharp deterioration in the situation for knowledge workers and possible unemployment at historically high levels.

Anthropic's economic team created the Scenario Explorer based on the technical report "Economic Scenarios for Transformative AI." The model views the economy as a collection of tasks performed by people, software, and machines, assessing which tasks AI can complement, automate, leave unchanged, or create from scratch.

The company outlined three scenarios:

• Moderate—AI's impact will be roughly the same as the advent of the internet. The economy will see an additional productivity boost, but it will remain within historical norms;
• Significant—by 2030, AI could perform half of all intellectual work, primarily autonomously, although companies will not use it for all such tasks. The economy will grow approximately twice as fast as normal rates;
• Extreme—AI will become more productive than humans in the vast majority of intellectual tasks and will perform nearly all such work autonomously. Implementing this scenario will likely require systems with recursive self-improvement and rapid AI deployment.

In the significant scenario, U.S. GDP could reach about $36.3 trillion in 2025 prices by 2030, while without considering AI's impact, it would be $33.5 trillion. In the extreme scenario, the figure could rise to $44.4 trillion.

At the same time, the emergence of AI creates new responsibilities, including quality checks for automated patient sorting or analyzing treatment plans proposed by the model.

According to Anthropic's estimates, the total value of all tasks performed in the U.S. economy by people, machines, and software exceeds $30 trillion per year. Therefore, AI's impact on specific tasks will directly affect GDP, employment, wages, and the distribution of income between workers and capital owners.

Anthropic notes that in all three scenarios, AI contributes to GDP growth, but the consequences for the labor market differ. In most cases, labor redistribution and unemployment levels will remain within historical norms, while in the case of extreme technological development, unemployment could reach record levels.

Workers in knowledge-related professions are at the highest risk, as AI may automate some of their tasks. Programmers, call center operators, and other specialists may need to retrain and transition to professions less prone to automation, but finding new jobs may take considerable time.

Average wages will rise in all three scenarios, but workers will benefit differently. In the significant scenario, incomes for knowledge workers will remain virtually unchanged, while in the extreme scenario, they could fall by more than 10% by 2030, whereas demand and wages in other professions will increase.

At the same time, large-scale automation may increase the share of income going to capital owners: currently, about 60 cents of every dollar of economic value goes to workers, while 40 cents go to capital. In the extreme scenario, the overall economic "pie" will significantly increase, but the share of workers will decrease, and total labor income will change little by 2030.

A survey by Anthropic among over 10,000 Americans in August showed that respondents' typical expectations are closest to the significant scenario: GDP by 2030 will be 10% higher than without AI, while unemployment is expected to rise to about 5%.