---
id: "entity-judge-william-alsup"
type: "entity"
source_timestamps: ["¶5", "¶14"]
tags: ["judiciary", "northern-district-california"]
related: ["entity-anthropic", "concept-fair-use-divergence", "concept-piracy-caveat", "quote-alsup-transformative", "quote-alsup-piracy"]
entityType: "person"
canonicalName: "William Alsup"
aliases: ["Judge William Alsup", "Judge Alsup"]
canonical_url: "cand.uscourts.gov"
speakers: ["Judge William Alsup"]
sources: ["tail2"]
isSpeakerEntity: true
---
## Segment 2 — tail2

## Article 126 — a126

# Judge William Alsup

A U.S. District Judge in the Northern District of California who presided over *Bartz v. Anthropic* ([[entity-anthropic-d2]]) and is one of the source's cited judicial voices.

In a pivotal, mixed June 23, 2025 decision he held that training LLMs on *lawfully acquired* copyrighted work is "exceedingly"/"spectacularly" transformative **fair use** (see [[quote-alsup-transformative]]) — while simultaneously establishing the **piracy caveat**: obtaining training data via piracy is "inherently, irredeemably infringing" and subject to statutory damages (see [[quote-alsup-piracy]], [[concept-piracy-caveat]]).

**Role in the source:** He anchors the AI-favorable pole of the fair-use divergence (see [[concept-fair-use-divergence]]) and is the origin of the piracy-caveat logic that drives [[claim-piracy-financial-risk]]. **Attributed quotes:** [[quote-alsup-transformative]], [[quote-alsup-piracy]].