OpenAI published internal economic research tracking how workers use AI outside their traditional job functions, mapping which novel activities become recurring behaviors versus one-time experiments.
The research matters because it moves past the standard productivity framing. Instead of asking whether AI makes workers faster at existing tasks, it asks what workers start doing that they never did before. That distinction shapes how companies should think about hiring, training, and role design.
Read the full piece for the methodology behind how OpenAI categorizes recurring versus experimental AI use, and which worker segments are showing the most behavioral change. The taxonomy they build here is the part worth scrutinizing.
[READ ORIGINAL →]