Methodology | AI Labor Index | AI Labor Index

1 Since both our work and Anthropic's is both based on the O*NET task atomic unit, we can join on their open-source dataset to display an "AI adoption score" for each occupation. We look forward to improving this metric with more usage data in the future.

A walkthrough of how we currently measure AI's potential impact on the global economy. AI is evolving rapidly, and we've avoided drawing firm conclusions from these findings. We recommend you do the same. Think of this as a snapshot: a starting point rather than a definitive guide. These point-in-time results, based on a specific sample of tasks and occupations, does not reflect every real-world scenario. Like any study, there are uncertainties, and different methods can yield different outcomes. If you'd like to provide feedback on an automation rating, complexity score, or other task…

related reading

• Introducing the Anthropic Economic Index \ Anthropicanthropic.com
• Labor market impacts of AI: A new measure and early evidence \ Anthropicanthropic.com
• Scenarios for our Economic Futureanthropic.com
• Economics and Transformative AI | Tom Cunningham – Tom Cunninghamtecunningham.github.io
• Toward an O*NET for AI R&Depochai.substack.com
• Measuring the performance of our models on real-world tasks | OpenAIopenai.com
• ai economy job impact.pdficeberg.mit.edu
• Remote Labor Index: Measuring AI Automation of Remote Workarxiv.org
• Focus areas for The Anthropic Instituteanthropic.com
• After Automation | Everyevery.to
• Anthropic Economic Index report: Cadencesanthropic.com
• [2510.25137] The Iceberg Index: Measuring Skills-centered Exposure in the AI Economyarxiv.org