This post exists as a stub because the question underneath it comes before every other page on this site. "AI" is doing enormous work in public conversation: the same word covers a spreadsheet formula, a coding assistant, and a prophecy about the end of work. A site that measures AI's effect on the labor market can't use the word that loosely — every score here depends on drawing the line somewhere specific.
The full post will set out, plainly, where that line sits: which systems the published data actually covers, and therefore what "AI exposure," "AI adoption" and "AI use at work" mean on this site — and what they don't mean. It will lean on the definitions already maintained in the glossary rather than inventing new ones.
Until then, the practical entry points are the glossary for the site's working definitions and How scores are computed for what each measure is built from.