A sharp editorial arguing that machine-use rights have become scholarly publishing's most under-resolved legal and operational problem. While the industry has workable routines for human readers (authenticate, license-check, serve, record), machine use — TDM access, governed corpora for AI analytics, workflow summarization tools — sits in a fragmented patchwork of institutional licenses, platform terms, API rules, and author agreements. The piece identifies the practical cost: product teams can't define permitted functionality, licensing teams struggle to package corpus access, and compliance infrastructure is only as strong as the underlying permissions. Not framed as resistance to AI, but as a structural gap that needs industry-wide resolution.

Source: Knowledge Speak