Legal Recognition Of Self-Modifying AI Code As A Form Of Creative AuthorshIP.

18 Mar 2026 --
0 Comments

1. Understanding Self-Modifying AI and Creative Authorship

Self-modifying AI refers to software systems capable of altering their own code, logic, or parameters in response to inputs or environmental changes. This raises unique questions in intellectual property law:

Who is the “author” of the work if the AI generates or modifies code autonomously?
Can copyright or patent protection apply to outputs created by AI without direct human intervention?
How do existing legal frameworks interpret authorship in these contexts?

Traditionally, copyright law requires a human author. Self-modifying AI challenges this principle because the AI can generate new creative expressions independently, raising the question of whether the AI itself could be recognized as a creator.

2. Key Legal Principles

Human authorship requirement – Copyright law generally presumes a human creator.
Originality & creativity – For protection, the work must reflect original creative expression.
Autonomy of AI – If the AI modifies itself and generates output, human control may be minimal.

Some jurisdictions (e.g., UK, EU, US) have begun exploring these issues, but courts are still cautious.

3. Case Laws Illustrating AI and Creative Authorship

(a) Naruto v. Slater (2018, US)

Facts: A macaque monkey named Naruto took selfies using a photographer’s camera. The photos went viral. PETA claimed the monkey should own copyright.
Holding: The US Copyright Office and courts rejected non-human authorship. Only humans can hold copyright.
Relevance to AI: Demonstrates the principle that authorship requires a human, posing challenges for AI-generated works, including self-modifying code.
Key Takeaway: Autonomous actors (animals, AI) are not recognized as authors under current US law.

(b) Thaler v. Commissioner of Patents (DABUS case, 2021, Australia)

Facts: Stephen Thaler filed patent applications listing an AI system, DABUS, as the inventor.
Holding: The Federal Court of Australia rejected the application because the law recognizes only human inventors.
Relevance to AI: Reinforces that even in inventive or creative works generated autonomously, current legal frameworks prioritize human authorship.
Key Takeaway: AI systems cannot be legally recognized as inventors or authors, but human oversight can secure rights.

(c) UK Copyright Office Guidance (2022) – Monkey Selfies Analogy Applied to AI

Facts: The UK Copyright Office issued guidance explicitly stating that for computer-generated works, copyright exists if there is human authorship in arranging, selecting, or programming the AI.
Key Implication: A human who creates the AI, writes the initial code, or sets parameters may claim copyright, even if the AI modifies itself.
Key Takeaway: Human involvement in guiding AI is critical to establish creative authorship legally.

(d) Feist Publications, Inc. v. Rural Telephone Service Co. (1991, US)

Facts: Feist compiled telephone directories; the question was whether mere factual compilation is copyrightable.
Holding: Courts held that mere collection of facts is not protected; originality is required.
Relevance to AI: Self-modifying AI generating code must display creativity or originality, not just mechanical output. A simple algorithm that modifies itself without creative input may not meet this threshold.

(e) European Parliament Resolution on AI & IP (2020)

Facts: The European Parliament examined AI-generated works, including autonomous music, art, and code.
Key Points:
- AI cannot be listed as an author.
- Human developers or operators may claim rights for their creative input in programming and configuring the AI.
- Encourages new frameworks for “AI-assisted authorship”.
Relevance: Indicates a move toward recognizing human-directed creativity in self-modifying AI systems, rather than granting AI independent legal personality.

(f) SAS Institute Inc. v. World Programming Ltd. (2012, EU)

Facts: World Programming developed software compatible with SAS’s programming language without copying source code.
Holding: Court of Justice of the European Union (CJEU) ruled that functionality and ideas themselves are not protected by copyright.
Relevance to AI: Suggests that even if AI autonomously generates functional code, copyright protection may only apply to original expression, not functional outcomes.

4. Implications for Self-Modifying AI Code

Authorship Attribution: Only humans can currently claim copyright for AI-generated code if they provided the initial framework, guidance, or direction.
Scope of Protection: Works entirely autonomous with minimal human intervention may fall outside traditional IP law.
Future Reforms: Legal scholars suggest introducing new IP categories like “AI-assisted works” or “computer-generated works” that can recognize contributions from self-modifying AI while preserving human oversight.

Summary Table of Cases and Principles

Case/Guideline	Jurisdiction	Principle	Relevance to AI Code
Naruto v. Slater	US	Only humans can own copyright	Autonomous AI cannot own copyright
Thaler v. Commissioner of Patents	Australia	Only humans recognized as inventors	AI cannot be inventor; human oversight needed
UK Copyright Office Guidance	UK	Copyright exists with human input	Human programming or intervention grants protection
Feist Publications v. Rural Telephone	US	Originality required for copyright	AI-generated code must show creativity
SAS Institute v. World Programming	EU	Functional aspects not protected	Only original expression, not functionality, can be protected
EU Parliament Resolution on AI & IP	EU	Supports human-directed authorship	Recognizes AI-assisted creativity under human supervision

Conclusion:
Currently, self-modifying AI cannot independently hold creative authorship under copyright or patent law. Legal protection exists only if humans provide the guiding creative input—whether in coding, training, or configuring the AI. Courts and regulatory bodies are increasingly recognizing AI-assisted authorship, but a fully autonomous AI as a legal author remains unrecognized.