Provenance After AI is a bridge packet by Lee Sharks that distinguishes several dimensions of provenance relevant to artificial intelligence.
Artifact provenance asks whether a file came from a declared source and records its edits. Licensing provenance asks under what permissions data entered a corpus or model. Semantic provenance asks whether an AI-generated synthesis preserves accountability to the people, works, traditions, and communities whose meaning it uses.
The packet does not claim that existing provenance systems failed at their intended purposes. It argues that large-scale synthesis creates a separate operational problem: meaning may retain valid file history and lawful corpus status while losing its intellectual lineage in the output.
Three provisional PER depths are proposed. Minimal provenance preserves author, source, date, and claim boundary. Conceptual provenance adds framework and tradition. Deep provenance includes social, ancestral, geographic, and futural obligation. These tiers remain research proposals requiring reliable annotation and validation.
The work is positioned as disciplinary clarification and extension. It acknowledges that provenance has always carried contextual meaning in archival and Indigenous traditions; the AI era changes the scale and technical conditions of possible erasure.
Referenced by 44 other entities in the archive. See the full Knowledge Graph for reference paths, or the primary record for the full deposit with reference details.