The scale sits between AL0 (no AI involvement), AL3 ("collaborates" — the AI does large chunks of work under close human direction), AL4 ("leads" — the AI completes most of a task end-to-end from a high-level prompt while a human supervises and approves before deployment, rather than staying actively engaged throughout), and AL5 ("autonomous" — no human in the loop at all). Anthropic adopted the scale in September 2026 to build its R&D Automation Index, an internal measure of how much of its own AI research and development work Claude performs, reporting Claude at 26% "leads" and zero measured work at AL5 as of August 2026.