Digitizing Knowledge: A Losing Chain of Custody

A recent court ruling has sparked debate about the consequences of digitizing physical knowledge. Anthropic, an AI research company, was involved in a case where they purchased millions of used print books, scanned their pages into training data, and discarded the original paper copies.

The court ruled that digitizing the books was fair use, citing that one replaced the other. However, this decision has raised concerns about the loss of provenance, the ability to prove what existed, when, in what form, and on whose attestation.

Provenance is crucial for verifying the authenticity of knowledge, especially as artificial intelligence (AI) democratizes access to information at an unprecedented scale. Without a public blockchain capable of handling unbounded scale, data provenance may be lost forever.