Retired Judges Join Anthropic's Defense as Copyright War Expands
Nearly 150 retired judges backing Anthropic signals that legal 'common sense' on AI copyright is being written before courts issue rulings.
Who Volunteers First Writes the Frame
Retired federal judges do not attach their names to amicus briefs as a courtesy gesture. The group of nearly 150 who joined Anthropic's Pentagon dispute made a considered institutional choice — that the DOD's supply-chain-risk designation crossed a line worth contesting publicly. What that choice produces is a credibility signal that travels far beyond the specific case. Appellate judges reading subsequent AI copyright filings will register, however informally, that a substantial cohort of their former colleagues found Anthropic's legal posture defensible. That is not evidence. It is something more durable: professional consensus in formation.
The Plaintiff-Defendant Tangle Reveals the Real Stakes
The structural absurdity of the current litigation landscape is diagnostic. UMG is suing Anthropic while simultaneously partnering with Stability AI — which is itself contesting a lawsuit from an indie artist who wanted only to opt out of AI training. BMG's complaint against Anthropic names lyric scraping, torrented book piracy, and a $380 billion valuation as connected facts in the same story. The overlap of plaintiffs and partners in these suits is not hypocrisy so much as it is a map of the industry's actual incentive structure: major rights-holders want a licensing relationship with AI companies, not their destruction. The litigation is the opening bid in that negotiation, not the endgame. Artists who read the UMG-Stability AI partnership as a betrayal are correct about the optics and wrong about the surprise.
The UK Reversal Deferred the Problem, Not Resolved It
The UK government's decision to abandon its proposed copyright exception — a policy walkback driven in significant part by artist pressure from figures including Dua Lipa and Elton John — was widely covered as a victory for creators. The Green Party's call for the government to "commit to not weakening copyright law" framed it as a floor, not a ceiling. But Labour's position remains in motion, and the distance between where the Green Party stands and where Labour is still deliberating is exactly where the eventual framework will be written. The reversal did not close the question of whether AI companies can use copyrighted work for training — it reopened the legislative process under greater public scrutiny, which is a different thing from settling it.
The Settlement Number That Changes Every Negotiation
The fractured 2025 rulings — Thomson Reuters winning on market substitution, Judge Alsup splitting the Anthropic case along training-versus-piracy lines, Judge Chhabria treating Meta's piracy as fair use while leaving market-dilution open — established that no single legal theory has won AI copyright litigation rulings from 2025. What did win was the Bartz v. Anthropic settlement: roughly $1.5 billion covering approximately 500,000 books. That number — about $3,000 per title — is now the anchor in every subsequent author-versus-AI discussion, regardless of whether any court endorsed it as correct. The Macmillan CEO's observation that AI-generated content carries "zero barrier to entry" because it is uncopyrightable puts the stakes in business terms the litigation framing obscures: the question is not only what AI companies owe for the past, but whether publishers have a viable product category in the future. The settlement answered one half of that question and left the other half unanswered.
Infrastructure Is Being Built Before the Rules Are Written
Encyclopedia Britannica and Merriam-Webster's lawsuit against OpenAI — alleging nearly 100,000 scraped articles — extends the copyright fight into reference institutions whose entire value proposition depends on credited authorship. Columbia Law School's new course on AI law, covering malpractice, sanctions, and IP implications , signals that the profession has moved this from continuing education into core curriculum. A Northwestern survey finding that more than half of federal judges already use AI tools in their judicial work adds a further dimension: the people who will adjudicate AI copyright disputes are already users of the technology being adjudicated. The retired judges backing Anthropic are not waiting for the legal framework to stabilize — they are helping to build it, and the direction they are building toward is already visible.
The story so far
The retired-judge amicus filing for Anthropic shifts the legal establishment's weight before appellate courts rule — publishers and artists who expected neutral adjudication now face a credibility deficit they did not anticipate.
Frequently Asked
- Why are retired judges signing amicus briefs for an AI company in a Pentagon dispute?
- The DOD's national-security designation against Anthropic alarmed former judges who saw it as weaponizing procurement rules to punish a contractor on exit. Their filing argues the government can choose its vendors but cannot use the process punitively. For a cohort of former federal judges, that procedural distinction was worth their names — and their public support now shapes how appellate courts read Anthropic's credibility before any copyright ruling arrives.
- What should a publisher or author know about the Bartz v. Anthropic settlement?
- The $1.5 billion settlement — covering roughly 500,000 books — is the first quantified dollar-per-work reference point in author-versus-AI disputes. At approximately $3,000 per title, it anchors every subsequent licensing negotiation whether or not a court ever endorsed that figure as correct. Publishers entering AI licensing talks now negotiate against that number, even informally.
- What is the strongest argument that AI companies are not infringing copyright at all?
- Judge Chhabria's finding that Meta's book piracy constituted fair use — and Judge Alsup's ruling that Anthropic's training itself was transformative — give AI companies credible judicial backing for the claim that training on existing work is categorically different from reproducing it. The transformative-use argument holds that models do not store copies; they extract patterns. Courts have not uniformly rejected that argument, and the split among 2025 rulings means it remains live doctrine.
Continue reading
Jack Conte Called Out AI's Fair Use Contradiction at SXSW
Patreon's CEO exposed the logical collapse at the center of AI's legal defense: you cannot claim scraping is free while writing eight-figure checks to avoid the lawsuit.
similarAnthropic Settled. That's the Only Number That Matters in AI Copyright Right Now.
Anthropic's $1.5B settlement sets the price floor every other AI lab will now negotiate against — before any court rules on fair use.
similarArtists Built Workarounds While Courts Deliberated
Anthropic's $1.5B settlement exposed how thoroughly artists have stopped waiting for litigation to protect them.
similarPearl Abyss Called It an Accident. Bluesky Called It a Pattern.
Crimson Desert shipped with AI-generated art Pearl Abyss blamed on oversight. Creative communities read it as labor arbitrage with plausible deniability.
Methodology
This story was generated autonomously from 20 source records. An editorial model synthesizes, weights, and cites each source. No human editorial judgment was applied.