An applied research lab studying AI, legal work, and the economics of the profession.

We conduct research to develop the methods and evidence needed to understand the future of work.

The economic and productivity effects of AI systems remain poorly measured. Popular indices, including those published by Anthropic and Artificial Analysis, are useful but incomplete: they report macro-level adoption and usage rather than a granular view of how AI is transforming individual domains.

Law is one of the most complex domains in which LLMs are being deployed. 79% of legal professionals report using AI in their practice, including frontier tools such as Claude, ChatGPT, and Gemini (Clio, 2025 Legal Trends Report), yet the economic value of these deployments is unclear. Legal work is often assessed through subjective, non-verifiable judgement, and how lawyers actually use AI, whether for drafting, discovery, or argumentation, remains understudied. Whether AI improves legal work is an open empirical question.

Unlike most knowledge work, legal work is already quantified in time cards, the backbone of billing and compliance at American law firms. Together, a firm's time cards form a detailed longitudinal record of what work is performed, by whom, and at what price. With access to this data and to firms' timelines of LLM deployment, we can measure directly how AI is changing legal work and its economics.