Copyright Concerns For AI-Generated Creative Therapy Modules In Healthcare
1. Understanding AI-Generated Creative Therapy Modules
Creative therapy modules in healthcare are AI-generated materials designed to aid in mental health, occupational therapy, or rehabilitation. Examples include:
AI-generated art exercises for stress relief
Music therapy playlists tailored by AI
Storytelling or guided meditation scripts
Key copyright questions:
Who owns the rights to AI-generated modules?
Can AI outputs qualify as original works of authorship?
How do derivative works or AI-trained datasets impact IP rights?
2. Legal Framework
Globally, copyright laws generally recognize human authorship as essential. Relevant principles:
Originality and Authorship:
Works must be original intellectual creations of humans.
Purely AI-generated outputs without human input typically do not qualify.
Derivative Works:
AI modules trained on copyrighted therapy programs may create derivative works, requiring licenses.
Computer-Generated Works:
UK CDPA 1988 Section 9(3) states that the human directing a computer-generated work is the author.
Many jurisdictions analogously apply this principle.
Moral Rights:
In healthcare and therapy, attribution and integrity are important for human authors contributing content.
3. Core Copyright Concerns
Authorship Ambiguity:
Who is the copyright owner: AI programmer, healthcare professional, or institution?
Originality:
Modules generated purely by AI may lack sufficient human creative contribution.
Derivative and Infringing Works:
AI trained on copyrighted therapy manuals or art/music programs can produce infringing outputs.
Ownership and Licensing:
Contracts with AI providers must clearly assign rights and ensure compliance with copyright laws.
4. Key Case Laws
Although few cases directly address AI in healthcare, analogous copyright cases provide guidance.
4.1 Thaler v. DABUS (UK, EU, Australia 2021–2022)
Facts: AI “DABUS” created inventions; Stephen Thaler claimed IP.
Holding: Only humans can be recognized as authors or inventors.
Relevance:
AI-generated therapy modules cannot hold copyright; human therapists or programmers must contribute creatively.
4.2 Zarya of the Dawn (US Copyright Office, 2023)
Facts: Comic book partially AI-generated.
Holding: Copyright applies only to human-authored portions.
Relevance:
Human intervention in therapy modules—such as selecting exercises, modifying scripts, or curating content—is necessary to secure copyright.
4.3 Infopaq International A/S v. Danske Dagblades Forening (C-5/08, EU 2009)
Facts: Software extracted short phrases from newspapers.
Holding: Only works reflecting human intellectual creation qualify.
Relevance:
Human therapists must shape AI-generated content to meet originality standards.
Mechanical AI outputs without human editing are not copyrightable.
4.4 Naruto v. Slater (US, 2018)
Facts: Monkey took selfies; humans claimed copyright.
Holding: Non-humans cannot hold copyright.
Relevance:
Reinforces that AI cannot be considered an author of therapy modules.
4.5 Feist Publications, Inc. v. Rural Telephone Service Co. (US, 1991)
Facts: Compilation of phone listings.
Holding: Only original selection/arrangement is copyrightable.
Relevance:
Human selection and ordering of AI-generated therapy exercises create protectable works.
Sequence of activities, choice of music, or narrative structuring can qualify as creative contributions.
4.6 Computer-Generated Works – UK Copyright Law (CDPA 1988, Section 9(3))
Provision: The human arranging the creation of computer-generated works is deemed the author.
Relevance:
In AI therapy modules, the healthcare professional programming or curating AI output qualifies as the author.
5. Practical Implications for AI-Generated Therapy Modules
| Issue | Implication | Strategy |
|---|---|---|
| Authorship | AI cannot hold copyright | Human therapists or programmers must actively guide and edit content |
| Originality | Pure AI output may not be protectable | Ensure selection, arrangement, and adaptation by humans |
| Derivative Works | AI trained on copyrighted therapy materials may infringe | License source materials or use public domain resources |
| Ownership | Ambiguity with AI vendors | Include IP assignment in contracts |
| Compliance | Protect patient confidentiality | Ensure AI output and data comply with healthcare privacy laws |
6. Summary Table of Case Law Relevance
| Case | Jurisdiction | Principle | Relevance to AI Therapy Modules |
|---|---|---|---|
| Thaler/DABUS | UK/EU/Australia | Only humans can hold IP | Human therapists/programmers are copyright holders |
| Zarya of the Dawn | US | Human-edited portions protected | Human curation ensures copyrightable modules |
| Infopaq | EU | Intellectual creation requires human input | Human shaping of AI outputs needed |
| Naruto v. Slater | US | Non-humans cannot hold copyright | AI cannot be author |
| Feist v. Rural | US | Original selection/arrangement protectable | Ordering and selection of AI exercises can be protected |
| CDPA 1988 s9(3) | UK | Human arranging computer-generated work is author | Directing or editing AI qualifies as authorship |
7. Conclusion
AI alone cannot hold copyright over therapy modules in healthcare.
Human creative input is critical, whether in selecting exercises, curating music, arranging art tasks, or modifying scripts.
Derivative works may infringe if AI is trained on copyrighted therapy resources.
Contracts and licensing with AI vendors are essential to assign rights clearly.
Documentation of human contribution is key for both copyright and regulatory compliance.
RELATED Blog
comments