Copyright OwnershIP In AI-Assisted Art Restorations.
1. Copyright Ownership in Art Restorations
1.1 Traditional Art Restoration
Traditionally, art restorations involve human expertise to repair or enhance an artwork while preserving its original style and aesthetic. The restoration process typically includes:
Cleaning the painting
Repainting missing or damaged areas
Repairing physical damage (e.g., tears in the canvas)
Art restoration has been treated as derivative works under copyright law, meaning the restored version is often copyrightable if it involves creative input (like filling in missing parts of a painting).
1.2 AI-Assisted Art Restoration
AI systems now assist restorers by predicting the missing parts of a painting, reconstructing damaged portions, or enhancing the original details to improve the quality of the artwork. AI-based restoration tools include:
Pattern recognition to detect damaged areas.
Data-driven prediction of missing artwork portions.
Image synthesis to generate realistic restorations.
However, questions about who owns the restored artwork's copyright emerge when AI is involved. Is the AI itself the author, or does the human restorer retain authorship?
2. Key Legal Principles and Case Laws
2.1 Feist Publications v. Rural Telephone Service Co. (1991, U.S.)
Facts:
Feist copied telephone directory listings from Rural Telephone and created its own directory. Rural claimed copyright over its listings.
Issue:
Whether compilations of facts (without creativity) can be copyrighted.
Judgment:
The U.S. Supreme Court ruled that facts alone do not receive copyright protection, but creative expression can be copyrighted.
Minimal creativity is required for a work to be protected under copyright.
Relevance for AI-Assisted Art Restoration:
AI-generated restorations that simply reproduce or reconstruct missing areas without creative human input may fail to meet the originality requirement.
If the AI-generated portions of the artwork are purely mechanical or data-driven, they may not qualify for copyright protection unless a human restorer’s creativity is involved in the final result.
2.2 Burrow-Giles Lithographic Co. v. Sarony (1884, U.S.)
Facts:
Photographer Sarony took a portrait of Oscar Wilde, and the photograph was copied without permission. The defendant argued that photographs (which are created mechanically) cannot be copyrighted.
Judgment:
The U.S. Supreme Court ruled that photographs are copyrightable if the photographer applies creative control.
The photographer’s creative decisions about composition, lighting, and framing were sufficient to grant copyright protection.
Relevance for AI-Assisted Art Restoration:
AI may assist in creating a restored version of an artwork, but human input in determining what needs restoration, how to handle missing elements, and applying creative judgment is necessary for copyright protection.
The human restorer’s creative judgment in selecting restoration techniques, rather than the mechanical AI process, would likely qualify for copyright protection.
2.3 Thaler v. Perlmutter (2023, U.S.)
Facts:
Stephen Thaler created an AI system, Creativity Machine, which generated artwork. Thaler tried to register the artwork with the AI listed as the author.
Issue:
Can AI be considered the author of a creative work?
Judgment:
The U.S. District Court held that AI cannot be the author of a work.
Copyright requires human authorship, and AI-generated works without human creative input cannot be protected under copyright law.
Relevance for AI-Assisted Art Restoration:
If AI is solely responsible for the restoration, with no human input, the restored artwork may not qualify for copyright protection.
However, if humans make the final decisions, selecting how the restoration should be performed or interpreting the results, they may claim copyright in the restored work.
2.4 Eastern Book Company v. D.B. Modak (2008, India)
Facts:
The Eastern Book Company published a compilation of legal cases, and D.B. Modak copied these cases.
Issue:
Whether a compilation of legal judgments can be copyrighted, even if the judgments themselves are in the public domain.
Judgment:
The Indian Supreme Court held that compilations requiring skill, judgment, and creativity could be protected.
Even though the legal cases themselves were in the public domain, the editor’s creativity in selecting, organizing, and presenting the content warranted copyright protection.
Relevance for AI-Assisted Art Restoration:
Similar to the compilation of legal works, if a human restorer applies creative skill to the AI-assisted restoration process (such as deciding how to restore missing portions based on their judgment of the artwork’s historical style), the restored artwork may qualify for copyright protection.
The AI-assisted restoration would be treated as a derivative work where the human contribution is copyrightable.
2.5 Nova Productions Ltd v. Mazooma Games Ltd (2007, UK)
Facts:
The case concerned the ownership of computer-generated graphics in video games. The defendant claimed ownership over graphics generated during gameplay.
Judgment:
The Court of Appeal held that the authorship of computer-generated works lies with the person who arranged and directed the creation of the work, rather than the machine that generated it.
Relevance for AI-Assisted Art Restoration:
In the context of AI-assisted art restoration, the person who curates the restoration (the human restorer or curator) may hold copyright in the restored artwork.
The AI is seen as a tool rather than an author, and the human curator’s creative decisions play a central role in determining the final restored version of the artwork.
2.6 Fox Film Corp. v. Doyal (1930, U.S.)
Facts:
Fox Film sought to claim copyright in an adaptation of a public domain play.
Judgment:
The U.S. Supreme Court ruled that derivative works (such as adaptations) could be copyrighted if they demonstrated sufficient original creativity.
Relevance for AI-Assisted Art Restoration:
Restoring a damaged artwork can be viewed as creating a derivative work.
If the AI-assisted restoration introduces new, creative elements that were missing in the original artwork, the restorer’s creative input can warrant copyright protection, as long as the restoration demonstrates originality in the reconstruction or enhancement.
3. Legal Framework and Practical Implications for AI-Assisted Art Restorations
1. Who Owns Copyright in AI-Assisted Art Restorations?
AI as a tool: AI is viewed as a tool or instrument in the restoration process, rather than the creator.
Human contribution: The restorer or curator who provides creative input, such as selecting the restoration technique, interpreting the artwork's original style, or making final adjustments, is likely to hold copyright in the restored work.
2. Derivative Works and Originality
If an artwork is restored by AI but retains a high degree of human creativity (e.g., in choosing colors, forms, and techniques), the final restored artwork may qualify as a derivative work with copyright protection for the human input.
3. Compilation of Public Domain Artworks
Like the Eastern Book Company case, if AI is used to reconstruct fragments of a public domain artwork, but the human restorer’s input organizes and presents these fragments in a creative way, the restoration may be copyrightable.
4. Summary of Case Laws and Key Takeaways
| Case | Key Principle | Relevance to AI-Assisted Art Restoration |
|---|---|---|
| Feist Publications v. Rural Telephone | Originality is required for copyright | Purely mechanical AI restorations may not qualify for copyright unless human creativity is involved |
| Burrow-Giles Lithographic Co. v. Sarony | Creative decisions make the work copyrightable | AI restorations need human judgment to apply creative control for copyright |
| Thaler v. Perlmutter | AI cannot be an author | AI-generated restorations without human involvement are not copyrightable |
| Eastern Book Co. v. D.B. Modak | Compilations requiring creativity can be protected | Human editorial decisions in restoring art may be copyrightable |
| Nova Productions Ltd v. Mazooma Games | Human guidance in computer-generated works determines authorship | Human curators/restorers can claim copyright in AI-assisted restorations |
| Fox Film Corp. v. Doyal | Derivative works can be protected with originality | Restoring artwork with creative human input can create a copyrightable derivative work |
Final Thoughts:
AI-assisted art restoration requires human creativity to qualify for copyright protection.
The AI system serves as a tool, and the restorer’s judgment in using AI to fill in missing details, recreate colors, or suggest missing elements is what grants the restored artwork copyright protection.
Derivative works from public domain art can be copyrightable if the restoration demonstrates originality in the process.
RELATED Blog
comments