📘 1. Introduction: What Is Algorithmic Authorship?

Algorithmic authorship refers to creative works—stories, poems, novels, essays—produced by artificial intelligence (AI) or machine learning algorithms, with minimal or no direct human creative input.

Traditional copyright law:

Assumes human authorship.

Grants rights when a natural person creates an original work.

AI challenges this model:

AI can generate entire novels, poems, and screenplays.

Who (if anyone) should have copyright?

Can the AI itself be recognized as an author?

⚖️ 2. Core Legal Principles

🧩 a) Human Authorship Requirement

Most jurisdictions require a natural person for copyright. AI cannot currently own copyright.

🧩 b) Human Contribution

When humans meaningfully shape, edit, or guide the AI output, copyright may vest in the human author or programmer.

🧩 c) Contractual Attribution

Parties (developers, users, publishers) may contractually assign rights.

📚 3. Detailed Case Law Analyses

Below are seven key cases that deal with algorithmic authorship or related issues in AI‑generated works, including detailed reasoning and legal outcomes.

Case 1 — Naruto v. Slater (“Monkey Selfie”)

Jurisdiction: United States, Ninth Circuit
Year: 2018

🧠 Facts

A macaque monkey pressed the shutter and took a selfie using a photographer’s camera. The photographer, David Slater, claimed copyright over the photo.

📜 Legal Issue

Is a non‑human “creator” eligible for copyright protection?

⚖️ Holding

No. The U.S. Ninth Circuit held:

Only human beings can be authors for copyright.

Animals and non‑humans do not qualify.

📌 Significance for AI

This case is foundational for AI literature:

It confirms that non‑human entities cannot own copyright.

AI—as a non‑human “creator”—cannot hold authorship status under U.S. law.

📍 Implication

If a novel is generated entirely by an AI with no human creative input, there is no copyright owner unless a human exercise significant editorial control.

Case 2 — Thaler v. Vidal (DABUS Patents)

Jurisdiction: United States Patent and Trademark Office (USPTO) & Federal Courts
Years: 2019–2021

🧠 Facts

Dr. Stephen Thaler submitted patent applications naming his AI system (DABUS) as the inventor.

📜 Issue

Can an AI system be legally recognized as an “inventor”?

⚖️ Outcome

Both the USPTO and U.S. Federal Courts rejected the applications.

Inventorship requires a natural person.

Cited statutory definitions of “inventor”.

📌 Significance for AI Authorship

While this case deals with patents rather than copyrights, the principle is similar:

AI may perform creative or inventive acts.

But legal recognition of AI as creator is rejected without statutory change.

📍 Implication

By analogy, AI cannot be legally recognized as the author of literature under current U.S. law.

Case 3 — Feist Publications v. Rural Telephone Service

Jurisdiction: U.S. Supreme Court
Year: 1991

🧠 Facts

Feist used information from Rural Telephone’s directory to create its own, and Rural alleged copyright infringement.

📜 Issue

Are facts and compilations inherently protected?

⚖️ Holding

No. Only original selection or arrangement is protected.

📌 Significance for AI Literature

Data or prompts used by an AI are like “facts.”

Copyright depends on original expression — which must show human creativity.

📍 Implication

An AI‑generated text may not be protectable unless a human adds creative organization or transformation.

Case 4 — Copyright Office Guidance on AI‑Generated Works

Jurisdiction: United States Copyright Office
Year: 2022 (official policy)

🧠 Policy Position

The U.S. Copyright Office clarified:

Works generated entirely by AI without human creative input are not copyrightable.

Works with meaningful human authorship may be registered.

📌 Key Elements

The Office considers:

Did a human direct the creative process?

Did the human make expressive choices?

Was the AI simply a tool like a word processor?

📍 Impact on AI Authors

If an author:

provides prompts,

refines outputs,

edits drafts,
they may qualify as a human author.

Case 5 — SAS Institute v. World Programming Ltd.

Jurisdiction: European Union (Court of Justice)
Year: 2012

🧠 Facts

World Programming copied the functionality of SAS software and competed with it.

📜 Issue

Is functionality (ideas, methods) protected by copyright?

⚖️ Holding

No. Only code expression is protected; not functional output or methods.

📌 Relevance to AI Literature

The process and model (code, training algorithms) may have copyright protection.

The output text may not be protected if it’s deemed method or functional.

📍 Implication

Protecting AI literature depends on expression — not underlying algorithms.

Case 6 — UK High Court: Thaler v. Comptroller General of Patents

Jurisdiction: UK High Court of Justice
Year: 2021

🧠 Facts

Another proceeding involving DABUS as an inventor claim in the UK.

⚖️ Holding

The High Court ruled:

Patents must be granted to person(s).

AI cannot be legally named as inventor.

📌 Parallel for Authorship

This reinforces the idea that European laws also do not recognize AI as a legal author.

Case 7 — European Parliament & Council Position (European Copyright Directive)

Jurisdiction: European Union
Year: 2019–2021

🧠 Legal Trend

EU law treats computer programs and databases as protected:

But for literary or artistic works, human creativity is essential.

AI outputs only gain protection if human authorship can be identified.

📌 Application

For an AI‑written novel:

Only the human who guided the process (e.g., editing, prompts) can claim copyright.

🧠 4. Comparative Legal Approaches — What Courts Are Saying

Here’s how different systems handle AI authorship:

🇺🇸 United States

Human authorship is essential.

AI cannot hold copyright.

Meaningful editorial contribution by humans is necessary for protection.

🇪🇺 European Union

Similar position: creativity and expression must originate from humans.

Software (AI code) is protected, but literary outputs are not autonomously copyrightable.

🇬🇧 United Kingdom

AI cannot be a legal author.

If AI is used, human choices must shape output.

📌 5. When AI‑Generated Literature Can Be Protected

AI‑generated works can obtain copyright only if:

✅ 1. A Human Makes Creative Contributions

Examples:

Provides detailed prompts.

Curates AI output.

Makes edits transforming raw output.

Arranges sections of an AI draft into a final work.

In these cases, the human is the author, and the work is protectable.

✅ 2. Contractual Assignment of Rights

If a publisher commissions AI work, a contract can:

Assign rights to authors or publishers.

Establish ownership even if AI is heavily involved.

Contracts can override defaults (but cannot give rights to AI itself).

✅ 3. Collaborative Authorship

If an author uses AI as a tool (like Photoshop or a camera), and makes expressive choices, courts treat the result as human‑authored.

📊 6. Summary: Key Legal Takeaways

📍 Conclusion

Algorithmic authorship in AI‑generated literature is not legally recognized at present. Courts and legal authorities around the world require:

Human creativity

Human editorial control

Human expressive choices

Only then can copyright be granted.

AI is a tool — not a legal author.
The output of AI, standing alone without human input, is generally not protected by copyright.