Copyright Implications For AI-Generated Immersive Storytelling In Italian Cultural Heritage Sites.
📌 1. What Is “AI‑Generated Immersive Storytelling”?
In the context of Italian cultural heritage sites (e.g., the Colosseum, the Uffizi, Pompeii), AI‑generated immersive storytelling refers to:
Narratives, characters, scripts, or multimedia experiences created or adapted by artificial intelligence models (e.g., generative text and image models).
Story content that changes interactively based on visitor input, sensor feedback, or real‑time interaction.
Presentations delivered through AR/VR headsets, projection mapping, apps, or site installations.
📌 Core Copyright Questions
1️⃣ Who is the author of AI‑assisted or AI‑generated works?
Is it the human who programmed/trained the AI, the site owner, or just the machine?
2️⃣ Can AI‑produced narratives enjoy copyright protection?
Can they be protected under Italian/EU law?
3️⃣ What about derivative works and existing public domain heritage elements?
📌 2. Key Legal Principles Under Italian/EU Law
Before cases, these principles matter:
✅ Copyright requires human creative input — EU/Italian law does not recognize machines as authors.
✅ Derivative works must show originality, not just rearrangement of existing material.
✅ Cultural heritage elements belong to the public domain — Italy protects its historic monuments, so AI stories must not simply regurgitate public domain texts without creative expression.
📌 3. Landmark Cases & Their Implications
Below are seven major case law examples, explained in detail, with emphasis on their relevance to AI, creative authorship, and immersive storytelling.
🔹 Case 1: Naruto v. Slater (Monkey ‘Selfie’ Case)
Jurisdiction: U.S. federal courts
Core Rule: Non‑human entities cannot own copyright.
Facts: A macaque monkey took a selfie with a photographer’s camera. Monkey’s claim to own copyright was rejected.
Holding: Only human authorship can hold copyright.
Relevance:
Similarly, an AI that writes a story cannot itself be an author under current law.
Ownership must be assigned to a human or legal entity — e.g., museum curator, creative director, or programmer who contributed creative choices.
Key takeaway: AI output alone is not automatically protected unless tied to human creativity.
🔹 Case 2: Feist Publications v. Rural Telephone Service
Jurisdiction: U.S. Supreme Court
Issue: Can a compilation of facts be copyrighted?
Holding: Mere data without originality is not protected.
Relevance for AI Storytelling:
Cultural heritage is full of facts and public domain content. AI that outputs a factual description of Pompeii or Dante’s life must add original expressive content to be protectable.
Simply compiling historical facts via AI has no originality and is not protected.
Key doctrine: Only original expressions (style, selection, narrative voice) qualify for copyright.
🔹 Case 3: U.S. Copyright Office’s AI Guidance — “Thaler/DABUS” Decisions
Summary (multiple actions by the U.S. Copyright Office):
The U.S. Copyright Office refused to register works produced solely by AI without human creative control.
Relevance:
Even where AI plays a major role, human contribution must be identifiable — e.g., editing prompts, structure, thematic choices.
Application to Italy/EU:
The underlying rationale — that the human must make creative contributions — carries persuasive weight globally, including in EU/Italian copyright enforcement.
🔹 Case 4: Bridgeman Art Library v. Corel (1999)
Issue: Are exact photographic reproductions of public domain art eligible for copyright?
Holding: No. Exact reproductions of public domain works are not sufficiently original to be protected.
Relevance:
AI storytelling that replays or reproduces historical narratives without creative transformation will not be protected. E.g.:
❌ AI reading out existing public domain texts verbatim
✔️ AI creating an original fictional narrative involving real historical events and characters
🔹 Case 5: Rogers v. Koons (1992)
Issue: What constitutes a derivative work?
Holding: Jeff Koons’ sculpture based on another photographer’s photo was infringing because it was not sufficiently transformative.
Relevance:
AI‑generated immersive scripts that draw from existing copyrighted material (e.g., books, films) must be highly transformative to qualify as new, protected works.
Example:
A VR experience loosely inspired by Dante but with an original storyline and character choices is more defensible than a direct retelling of The Divine Comedy.
🔹 Case 6: Authors Guild v. Google (2015)
Issue: Is scanning massive texts for a searchable database fair use?
Holding: Yes — it was transformative.
Relevance:
If an AI model uses public domain heritage texts to generate new interpretations or contexts (e.g., emotional retellings, interactive layers), that can be seen as legal and creative transformation rather than infringement.
🔹 Case 7: Thaler v. Commissioner of Patents (DABUS), Australia (2021)
Issue: Can an AI be recognized as an inventor in patent law?
Holding: The Australian patent office recognized an AI as an inventor (rare globally).
Relevance for Immersive Platforms:
This case is not about copyright, but it shapes how jurisdictions view AI contributions.
If some parts of immersive storytelling systems are patented (e.g., adaptive narrative engines), then creators must clarify ownership and inventorship upfront.
🔹 Case 8: European Court of Justice (ECJ), “Infopaq” & “Painer” Cases (EU Originality Standards)
Infopaq (2009):
Protected portions of text must be an expression of the intellectual creation of the author.
Painer (2011):
Photographs require creative choices that reflect personality.
Relevance:
EU/Italian law requires creative choices by a human — narrative structure, selection of events, tone, themes — not just machine output.
📌 4. Key Takeaways for Italian Cultural Heritage Context
✅ Human Creative Involvement Is Essential
To secure copyright for AI‑generated immersive narratives:
Human must shape prompts, adapt content meaningfully
Provide editorial control, narrative direction, and creative decisions
Attribution to humans (curators, writers, directors) is vital
✅ Public Domain Works Can Be Used, But Only if Transformed
Italian cultural heritage (many texts and historical facts) is largely public domain, but:
✔️ Creative reinterpretations are protected
❌ Pure recitation of facts or public domain work is not
Example:
A VR story imagining how gladiators felt based on historical research can be protected.
Simply converting a public domain text into audio with AI isn’t.
✅ Derivative Works Must Add Originality
Under Rogers v. Koons, simply layering old work on new tech does not suffice.
✅ Ownership Agreements Are Critical
All contracts with AI developers, museums, artists, and site operators should:
Clarify ownership of AI tools and outputs
Establish who owns rights to interactive stories
Protect underlying algorithms or narrative engines (via trade secret/patent when possible)
📌 5. Practical Guidance for Creators at Italian Heritage Sites
| Issue | Legal Implication |
|---|---|
| AI generates stories | Only protectable if human creative judgment is evident |
| AI learns from public domain art | Public domain is free, but new expression must be original |
| Collaborative work with programmers | Agreements must assign rights clearly |
| Derivative content | Must be transformed beyond existing works |
| Use of heritage texts | Can be the basis for new narratives but not protected by themselves |
📌 6. Conclusion (Key Legal Insights)
✔️ Human authorship is required — AI can assist, but cannot independently own or generate copyrightable works.
✔️ Original expressive choices matter — narrative structure, tone, character arcs.
✔️ Public domain materials help but don’t guarantee protection.
✔️ Derivative content must be sufficiently transformative.
✔️ Clear contracts are needed to avoid disputes between museums, artists, and AI developers.
RELATED Blog
comments