Copyright Implications For AI-Generated Stories Based On Biometric Rhythm And Breath.
📌 1) Thaler v. Perlmutter / U.S. Copyright Office — AI Alone Cannot Be an Author
Court & Year: U.S. Court of Appeals (D.C. Circuit), 2025
Key Holding: A work created entirely by an AI system without meaningful human creative input is not eligible for copyright protection because copyright law requires human authorship.
Facts:
Dr. Stephen Thaler created an AI called the Creativity Machine, which generated an artwork titled “A Recent Entrance to Paradise.”
He applied to the U.S. Copyright Office, listing the AI as the sole author.
The Office rejected it because the work was generated autonomously by software with no human creative control.
Why It Matters:
The appeals court agreed with the Copyright Office that:
The word “author” in the U.S. Copyright Act refers only to a human being;
Purely AI-generated works lack the necessary originality from a human mind to be protected.
This means an AI-generated story based solely on biometric rhythm + breath with no significant human creative input would be uncopyrightable.
The owner of the AI or the biometric data provider cannot claim copyright unless there’s significant human artistic contribution.
Application to Biometric AI Stories:
If you feed breathing data into an AI and it outputs a narrative without significant creative framing or selection by a human author, courts are likely to say: “No copyright.” The underlying data or algorithmic generation alone do not make it a copyrighted literary work under current U.S. law.
📌 2) Bartz v. Anthropic — Fair Use in AI Training & Copyright Liability
Court: U.S. District Court (Northern District of California), Fair Use Ruling, 2025
Key Holding: Training an AI on copyrighted text can be fair use if certain conditions are met, but unauthorized use may still create liability.
Facts:
A group of authors (including Andrea Bartz) sued AI developer Anthropic, claiming the company’s models were trained on copyrighted books without authorization.
The court held training on books that Anthropic had legitimately purchased could qualify as fair use, likening the model to a human “learning” from existing works.
Why It Matters:
Courts are starting to treat large language models (LLMs) analogously to human learners in terms of fair use.
But how data was obtained matters: using pirated, unlicensed sources could still be actionable as infringement.
Application to Biometric AI Stories:
If your story-generation AI was trained on copyrighted narrative data without permission, and it outputs text incorporating that material too closely, the model developers (and possibly users) might face copyright claims. Using licensed or public-domain sources lessens this risk.
📌 3) Meta Platforms Copyright Lawsuit (Authors v. Meta) — Fair Use & Argument Quality
Court: U.S. District Court, 2025
Key Holding: Meta successfully defended against a copyright suit by authors who alleged unauthorized AI training — but the ruling was narrow and not a broad license.
Facts:
13 authors sued Meta, saying its AI training infringed their works.
The judge dismissed the lawsuit, not because AI training is necessarily lawful, but because the plaintiffs failed to build a record supporting their claims on key issues like copying and market harm.
Why It Matters:
Even where an AI is alleged to have infringed, outcomes can hinge on how the legal arguments are framed and on the specifics of how the model uses the data. It doesn’t mean all AI training is safe; it means courts are still sorting through these new issues carefully case by case.
Application to Biometric AI Stories:
If your biometric-based AI uses or reproduces copyrighted narrative text, you must be ready to show strong evidence of non-infringing use or qualify for fair use (transformative use, limited excerpts, no market harm, etc.).
📌 4) GEMA v. OpenAI (Germany) — AI Output Reproducing Protected Works
Court: Regional Court of Munich, Germany, 2025
Key Holding: An AI model can infringe copyright when it memorizes and reproduces copyrighted material without license.
Facts:
GEMA (German music rights society) sued OpenAI, alleging the LLM could output copyrighted song lyrics.
The German court ruled that if the AI regurgitates protected material, this is an act of infringement under German/EU law.
Why It Matters:
Unlike the U.S. focus on human authorship, European law can treat AI reproduction of copyrighted content seen in training data as infringement, regardless of whether a human is involved.
Application to Biometric AI Stories:
If your system uses narrative or copyrighted text as part of its reasoning and reproduces it verbatim, it could be infringing under EU/Indian/Eastern Hemisphere laws that emphasize reproduction rather than human authorship.
📌 5) Authors Guild, Inc. v. HathiTrust — Fair Use for Machine Processing
Court: U.S. 2nd Circuit Court of Appeals (HathiTrust), 2014
Key Holding: Systematic copying of copyrighted books for search and accessibility can be fair use, even if full copies are stored.
Facts:
Authors sued a digital library for copying books scanned by Google.
The court found that transformative uses like search indexing were fair use.
Why It Matters:
This established early that technological uses of copyrighted material (even full copies) could be fair use if the public benefit and transformation outweigh market harm.
Application to Biometric AI Stories:
If your biometric story system indexes or analyzes copyrighted works non-expressively and only uses outputs as prompts, you might argue a transformative use — but this doesn’t apply if the final output is a spatially similar text readable to humans.
🔍 Additional Foundational Principles from Case Law
âť— Human Authorship Requirement
Across U.S. cases and Copyright Office interpretations (e.g., the Thaler decision), the consistent theme is:
Copyright attaches only to works with substantial human creative input.
An AI system generating text from biometric data without meaningful human editorial involvement will likely fail this test.
âť— Idea vs. Expression
Cases like Ho v. Taflove (scientific data not copyrightable) underline that ideas, facts, or data (e.g., biometric rhythms) themselves aren’t protected — only the expression of creative concepts is.
đź“Ś Practical Implications for Biometric AI Stories
1. Output Without Human Input = Likely No Copyright
A fully autonomous AI story based on biometric rhythm and breath with no human creative “touch”—just data input and model output—is unlikely to be copyrighted under current U.S. law (and many other jurisdictions).
2. Human Creative Input Can Salvage Protection
If you as author shape, select, revise, and significantly contribute original text, you could claim copyright in the creative parts — even if parts came from AI. The human contribution must be more than a simple prompt.
3. Training Data Matters
Using copyrighted texts to train your model increases risk of infringement if the AI reproduces them. Licensing training data or using public domain sources lowers risk.
4. International Variations
EU and Indian law may treat generative AI output differently — some focus on reproduction rather than human authorship, so you must check local statutes and court decisions.
🧠Summary
| Aspect | Likely Legal Position |
|---|---|
| Is purely AI-generated biometric story copyrightable? | Generally no (no human authorship). |
| Is a human-edited AI story protectable? | Yes, if original human creative contribution is clear. |
| Can training an AI on copyrighted works be infringement? | Yes, unless fair use or license applies. |
| Can regression (reproduction) of copyrighted text by AI be infringing? | Yes, especially in Europe. |
| Are biometric data themselves protected? | No, data not copyrightable, but expression based on data may be. |
RELATED Blog
comments