A group of publishers and authors sued Google on July 14, 2026. The class action accuses the company of training its Gemini models on copyrighted works without permission, according to TechCrunch. The plaintiffs include Hachette, Cengage, Elsevier, author Scott Turow, and the group S.C.R.I.B.E.
Freelance writers, editors, illustrators, and course creators aren’t named parties. But they have a direct stake. The courts are now deciding whether companies can ingest the words you sell for free.
What The Lawsuit Actually Alleges
The complaint claims Google trained Gemini on books it obtained through scope-limited programs. Those programs included Google Books and titles uploaded to the Google Play Store. Publishers had supplied those works for a different purpose: making books searchable in snippet form, not feeding them into an AI model.
The plaintiffs also allege something more serious. They say Google intentionally removed or altered copyright management information to hide that its Gemini models trained on what the filing calls stolen materials. That claim matters because it goes beyond fair use. It reaches into a separate part of copyright law entirely.
The suit cites an internal Google document. That document allegedly warned that using copyrighted books for AI training could be highly problematic for the company. It reportedly estimated potential fines in the tens to hundreds of billions of dollars. Google did not immediately respond to a request for comment.
Why This Matters For Self-Employed Creators
Two early California decisions already shaped this debate. Courts in the Anthropic and Meta cases found that training on copyrighted books was likely fair use. This new case lands in the Southern District of New York instead. That means a different court and a different judge will weigh in.
The money question remains unsettled too. Anthropic paid $1.5 billion over the works it pirated — the largest payout in US copyright law history. Roughly half a million writers became eligible for payments of at least $3,000 each. Many authors opted out of that settlement so they could pursue their own claims.
For a solo creator, the practical takeaway is simple. Licensing, not litigation, is where the real leverage will land. Courts keep splitting on fair use, and that split gives AI companies a growing incentive to buy access instead of assuming it.
What Self-Employed Creators Should Do Next
Start with your contracts. Check whether your client agreements grant broad, perpetual, sublicensable rights. That kind of language can let a client feed your work into a model without ever owing you more.
Add an explicit AI training clause to new agreements. You can permit it, price it separately, or prohibit it outright. Silence is the one option that reliably works against you.
Keep records of what you published and where. If a settlement or licensing pool emerges later, eligibility will hinge on documentation. Reconstructing a decade of bylines after the fact is painful work — start the habit now.
What To Watch Next
Watch whether the New York court diverges from the California rulings on fair use. A split between circuits usually pushes a question toward the Supreme Court or Congress. US copyright law hasn’t seen a meaningful update since before the internet existed.
Also watch how licensing deals reprice creative work. AI is already rewriting freelance pay as value pricing takes over. Formal training licenses would add another revenue line — one solo creators will need to learn how to negotiate.
Photo by Firmbee.com: Unsplash
