What Every Author Should Know About AI Training
AI training and author rights have become one of the most urgent issues in modern publishing. As artificial intelligence systems become increasingly capable of writing, summarizing, and even mimicking human style, authors are asking: Where did these models get their data — and was it used with permission? The recent Anthropic settlement highlighted that copyrighted books have sometimes been used to train large language models (LLMs) without the authors’ consent.
At Atmosphere Press, we believe every author — whether you publish with us, another press, or independently — deserves full control over how your work is used. This article explains our position on AI training of copyrighted books, why this issue matters for all authors, what recent legal developments mean, and practical steps you can take to protect your creative work.
TABLE OF CONTENTS
What Happened: The Anthropic Settlement in Brief
AI Training and Author Rights
Legal Note About Copyright Registration
Atmosphere Press’s Policy on AI Training
Safeguarding Your Author Rights
FAQ
What Happened: The Anthropic Settlement in Brief
A major class action against Anthropic, the maker of the Claude models, has led to a proposed $1.5 billion settlement. The lawsuit alleged that Anthropic acquired and used copyrighted books without permission to train its AI models.
As reported by Reuters and Publishers Weekly, the settlement could provide approximately $3,000 per title for eligible rights-holders. The settlement covers past use, not future use, and includes a process to determine which works were included. While final approval is pending, the case serves as a landmark example of how copyright infringement in AI training is being addressed.
There are likely to be other similar lawsuits, so keep an eye out. We’re in an evolving legal landscape, that’s for sure!
AI Training and Author Rights
Artificial intelligence is reshaping the creative world, but at its core, the issue of AI training and author rights comes down to ownership and consent. Every author deserves to control how their words are used — whether that’s for reading, quotation, adaptation, or AI training.
Cases like the Anthropic settlement highlight that technology can outpace traditional copyright frameworks. Authors must be proactive in protecting their work to ensure creative labor is respected and intellectual property remains under their control.
1. Consent and Control
Authors — and the publishers they contract with — should decide whether their books can be used as training data for commercial AI models. One way to assert this right clearly is through a copyright-page notice. For our Atmosphere Press authors, we include the following notice in every book:
“Unless author permission is expressly granted, any use of this publication to ‘train’ artificial intelligence (AI) technologies to generate text is prohibited.”
Even if you’re published elsewhere or self-publishing, you can implement a similar statement to protect your work and communicate consent clearly!
2. Practical Consequences
The Anthropic settlement demonstrates that rights-holders may be entitled to recoveries if their work is used without authorization. Eligibility for claims often depends on proper documentation, registration, and clear proof of ownership. This underscores the importance of maintaining accurate records and considering copyright registration, which can strengthen an author’s position in disputes.
3. Broader Implications for All Authors
Understanding and asserting your rights matters, whether you’re an established author, newly published, or just starting. AI models continue to be trained on public and private texts, so proactively clarifying consent protects your creative output, ensures control over its use, and signals to the industry that authors’ rights are non-negotiable.
Legal Note About Copyright Registration
Under U.S. law, your work is copyrighted as soon as it’s written. In that way, filing for additional copyright protection has historically not been necessary. However, formal registration with the U.S. Copyright Office does provides advantages:
✦ Required for filing lawsuits or collecting statutory damages in most cases.
✦ Creates an official record of ownership and date of authorship.
✦ Can make you eligible for certain settlements and claims.
Because Atmosphere Press authors retain their own rights to their work, Atmosphere Press does not file formal copyright registrations for authors. With these latest legal developments, though, we do encourage every author to register independently! You can start at copyright.gov/registration; the process is simple, inexpensive, and provides significant protection.
Atmosphere Press’s Policy on AI Training
Atmosphere Press does not allow book content to be used for LLM training without explicit permission. Every title now includes the following notice on the copyright page:
“Unless author permission is expressly granted, any use of this publication to ‘train’ artificial intelligence (AI) technologies to generate text is prohibited.”
This protects our authors’ creative work and makes our stance clear to AI companies.
Like many forward-thinking publishers, Atmosphere Press uses AI responsibly — for manuscript analysis, comparative research, and marketing — but we never feed author manuscripts or book files into external training datasets.
Our philosophy is simple: AI can support publishing, but it should never exploit creativity.
Safeguarding Your Author Rights
1. Register your copyright
Visit the U.S. Copyright Office registration page. It’s a small step that provides significant protection and can affect whether you’re eligible for settlements or enforcement.
2. Add a “no AI training” notice
Even if you self-publish, you can add a short notice on your copyright page, such as:
“No portion of this work may be used to train large language models or similar AI systems without written permission.”
This signals your intent and strengthens your rights.
3. Keep records
Maintain copies of contracts, ISBN assignments, proof of publication, and final manuscript files. These documents help prove authorship and publication dates if issues arise.
4. Stay informed
Follow updates from trusted sources like The Authors Guild, Publishers Weekly, and the U.S. Copyright Office.
5. Reach out for guidance
If you’re an Atmosphere Press author (or thinking of becoming one), our editorial team can confirm your copyright-page wording and point you toward helpful registration resources!
FAQ: Authors’ Rights, AI, and Atmosphere Press’s Position
Does this mean Atmosphere Press is anti-AI?
Not at all. Atmosphere Press uses AI tools responsibly for data insights and workflow efficiency. Our issue is with unauthorized use of creative work, not the technology itself.
What exactly counts as “training” an AI model?
“Training” means feeding vast amounts of text into algorithms so they can learn how language works. If your book was included in those datasets without permission, that’s a potential copyright issue.
Can I choose to let my book be used for AI training?
Yes. Consent is key. If an author wishes to license their work to an AI company under specific terms, that’s their choice.
Does Atmosphere Press register copyrights for authors?
No. Authors own their copyrights and are encouraged to register them independently for the greatest protection.
Could my book already be part of an AI dataset?
Possibly. The Anthropic settlement includes a process to identify affected works. Keep an eye on updates from the Authors Guild or the court administrator for how to check.
Have any Atmosphere authors been affected by AI training lawsuits?
To our knowledge, no books published by Atmosphere Press have been directly impacted by AI training lawsuits.
Will I automatically receive money from the settlement?
No. Authors generally must file claims once the window opens. Registration and documentation will help you prove eligibility.
Does the “no LLM training” notice have legal weight?
Yes — it makes your prohibition explicit, which strengthens your position if your work is used without permission.
Can self-published authors use the same wording?
Absolutely. You can — and should — include your own “no AI training” language on your copyright page.
What if I find out my work was already used?
Monitor ongoing cases, file claims if eligible, and consider consulting an attorney. The Authors Guild often provides resources to guide that process.
Will Atmosphere Press ever share data with AI companies?
No. Atmosphere Press never provides book or manuscript data to train external AI models.
AI Training and Author Rights: Atmosphere Press’s Stance
Protecting authors’ rights is central to responsible publishing. Whether you publish with Atmosphere Press, another press, or independently, the principles are the same: control, consent, and respect for creative ownership. Atmosphere Press will continue using AI in ways that empower authors — not exploit them — and advocate for transparent, ethical innovation. We are committed to helping every writer navigate this evolving landscape!
Erin K. Larson-Burnett, Production Manager at Atmosphere Press (submit your manuscript here!), is a born-and-raised Southerner currently living in Katy, Texas, with her husband and their small domestic zoo. She is an avid ink drinker who lives and breathes books—during the day, she works remotely with authors around the world, honing and perfecting books published through Atmosphere Press. By night, she crafts her own stories…or at least tries to. The Bear & the Rose is her debut novel.
Atmosphere Press is a selective hybrid publisher founded in 2015 on the principles of Honesty, Transparency, Professionalism, Kindness, and Making Your Book Awesome. Our books have won dozens of awards and sold tens of thousands of copies. If you’re interested in learning more, or seeking publication for your own work, please explore the links below.