Copyright Issues In AI-Assisted Art Restoration Projects.
1. Introduction: AI-Assisted Art Restoration and Copyright
AI-assisted art restoration involves using algorithms to reconstruct, clean, or complete damaged artworks. Examples include restoring missing parts of paintings, recreating faded colors, or generating virtual versions of lost sculptures. Key legal questions include:
Who owns the copyright in AI-restored art?
If AI performs the bulk of restoration with minimal human input, can anyone claim copyright?
Derivative works
Restored artwork may be considered a derivative of the original, raising infringement issues.
Moral rights of the original artist
Many jurisdictions protect an artist’s right to integrity and attribution, which AI restoration may affect.
Public domain vs. copyrighted works
Restoration of public domain art has fewer legal hurdles, but copyrighted artworks require permission.
2. Key Copyright Principles
A. Human Authorship
Most jurisdictions require human creativity for copyright protection.
AI-assisted restoration with minimal human input may not create a new copyright.
B. Derivative Works
Copyright law generally protects derivative works.
Restoring a copyrighted painting without permission may constitute infringement, even if AI is used.
C. Moral Rights
Artists can object to modifications that distort or mutilate their work.
AI restoration must consider integrity and attribution rights.
3. Relevant Case Laws
Here are seven significant cases illustrating AI, derivative works, and art restoration issues:
Case 1: Naruto v. Slater (2018, 881 F.3d 983, 9th Cir., USA)
Facts:
A monkey took a selfie, raising the question of non-human authorship.
Ruling:
Only humans can hold copyright; non-human creations cannot.
Relevance:
AI-assisted restoration alone, without significant human creative input, may not produce a copyrightable work.
Case 2: Bridgeman Art Library v. Corel Corp. (1999, 36 F.Supp.2d 191, SDNY)
Facts:
Photographic reproductions of public domain artworks were claimed as copyrighted.
Ruling:
Exact reproductions lack originality; no copyright exists.
Relevance:
AI restorations that only replicate existing artwork may not qualify as new works, especially for public domain art.
Case 3: Community for Creative Non-Violence v. Reid (1989, 490 U.S. 730)
Facts:
A sculptor created a work as an independent contractor.
Ruling:
Copyright resides with the author unless it is a “work for hire.”
Relevance:
In AI-assisted restoration, human operators’ contributions (choices about color, reconstruction, or repair) may determine who owns copyright.
Case 4: Authors Guild v. Google, Inc. (2015, 804 F.3d 202, 2nd Cir., USA)
Facts:
Google scanned copyrighted books for a searchable database.
Ruling:
Court recognized transformative use and fair use.
Relevance:
AI restoration can be seen as transformative, especially when used for educational or research purposes. Commercial restorations of copyrighted works without permission may still infringe.
Case 5: TVEyes, Inc. v. Fox News Network (2018, 883 F.3d 169, 2nd Cir., USA)
Facts:
TVEyes scanned broadcast content for search and analysis.
Ruling:
Transformative use can qualify as fair use, but commercial exploitation may require licenses.
Relevance:
AI-assisted restoration transforms content, but commercialization (selling restored works) may require permission from copyright holders.
Case 6: Moral Rights Case – Snow v. Eaton Centre Ltd. (1982, Ont. H.C.)
Facts:
Sculptor Michael Snow’s installation had ribbons tied around it without his consent.
Ruling:
Court upheld moral rights, protecting the artist’s integrity.
Relevance:
AI restoration altering an artwork’s appearance could violate moral rights if the original artist’s integrity is compromised.
Case 7: US Copyright Office AI Policy (2023 Update)
Facts:
Clarified that works generated entirely by AI without human creative input are not copyrightable.
Relevance:
AI-assisted restorations require significant human input to create copyrightable works. Pure AI automation cannot claim copyright.
Case 8: Feist Publications, Inc. v. Rural Telephone Service Co. (1991, 499 U.S. 340)
Facts:
Dispute over copyright in telephone directory listings.
Ruling:
Copyright requires originality and minimal creativity; mere effort is insufficient.
Relevance:
AI restoration must include creative decisions, not just technical reconstruction, to qualify for copyright.
4. Practical Implications for AI-Assisted Art Restoration
Human Oversight is Essential
Ensure curators, restorers, or artists make significant creative decisions on AI output.
Derivative Works Require Permission
AI restoration of copyrighted artworks without authorization may constitute infringement.
Respect Moral Rights
Even with permission, restoration must preserve integrity and attribution of the original artist.
Public Domain Works
AI restoration of public domain art is safer for copyright but still requires careful attribution and transparency.
Documentation
Keep records of human creative input and AI decision-making for legal clarity.
5. Conclusion
AI-assisted art restoration intersects with copyright law in nuanced ways:
Pure AI restoration = likely no copyright.
Human-guided restoration = copyrightable if there’s originality.
Restoring copyrighted works without permission = potential infringement, regardless of AI.
Moral rights must be respected, even if AI technically performs the restoration.
Case law (Bridgeman, Naruto, Reid, Authors Guild, Snow, Feist) emphasizes human authorship, originality, transformation, and integrity, which are all critical for legally compliant AI-assisted restoration projects.
RELATED Blog
comments