Ipr In Licensing AI-Generated Software.
IPR in Licensing AI-Generated Software
Detailed Explanation with Case Laws
1. Introduction: AI-Generated Software and IPR
AI-generated software refers to computer programs, code, or software components that are created wholly or partially by artificial intelligence systems, such as:
Code generated by large language models
Auto-generated algorithms
AI-assisted debugging or optimization tools
Self-learning software systems
The core legal challenge is that traditional IPR laws are human-centric, while AI systems:
Do not have legal personality
Do not qualify as “authors” or “inventors”
Often rely on training data that includes copyrighted code
Licensing AI-generated software raises questions about:
Ownership
Authorship
Patent inventorship
Copyright validity
Liability and infringement
2. Types of IPR Relevant to AI-Generated Software
(a) Copyright
Protects source code, object code, and software architecture
Requires human authorship in most jurisdictions
(b) Patents
Protect technical solutions implemented by software
Require a human inventor
(c) Trade Secrets
Protect AI models, training processes, and proprietary algorithms
(d) Licensing
Licensing determines:
Who may use AI-generated software
Whether outputs are proprietary or open-source
Allocation of liability for infringement
3. Licensing Models for AI-Generated Software
Provider-centric licensing
AI developer owns outputs; users get limited rights
User-centric licensing
User owns AI-generated code outputs
Shared or conditional licensing
Ownership depends on level of human involvement
Open-source influenced licensing
AI outputs subject to open-source compliance obligations
4. Landmark Case Laws
Case 1: Thaler v. Comptroller-General of Patents (UK, 2021–2023)
Facts
Stephen Thaler filed patent applications naming an AI system (DABUS) as the inventor.
Claimed that AI independently generated the invention.
Legal Issue
Can an AI system be recognized as an inventor under patent law?
Who owns AI-generated inventions for licensing purposes?
Decision
UK courts rejected the application.
Held that only a natural person can be an inventor.
Reasoning
Patent law assumes legal rights and duties that only humans can hold.
AI cannot transfer ownership rights.
Significance
Licensing AI-generated software patents requires human attribution.
Reinforces that AI output must be legally linked to a human.
Case 2: Thaler v. United States Patent and Trademark Office (USA, 2022)
Facts
Thaler sought patent protection in the US for AI-generated inventions.
USPTO rejected the applications.
Legal Issue
Whether US patent law allows non-human inventors.
Decision
US courts affirmed USPTO’s rejection.
Reasoning
Patent Act consistently refers to inventors as individuals.
Congress did not intend to include AI.
Significance
AI-generated software patents must be licensed through human inventors or assignees.
AI cannot be a direct licensor or licensee.
Case 3: Naruto v. Slater (Monkey Selfie Case – Analogy, USA, 2018)
(Though not AI, it is legally influential)
Facts
A monkey took photographs; activists claimed copyright on its behalf.
Legal Issue
Can non-humans own copyright?
Decision
Court held non-humans cannot hold copyright.
Significance
Frequently cited by courts and scholars to deny copyright in purely AI-generated works.
Impacts licensing: no copyright = no exclusive license.
Case 4: US Copyright Office – Zarya of the Dawn Decision (2023)
Facts
Applicant sought copyright over a graphic novel created using AI tools.
Claimed ownership of AI-generated images.
Legal Issue
Is AI-generated content copyrightable?
Decision
Copyright granted only for human-authored elements.
AI-generated parts were excluded.
Reasoning
Copyright protects human creative choices, not machine outputs.
Significance
Licensing AI-generated software requires:
Proof of human creative control
Clear attribution in license agreements
Case 5: Oracle America v. Google (USA, 2016–2021)
Facts
Google used Java APIs in Android.
Oracle claimed copyright infringement.
Legal Issue
Can software interfaces and reused code be protected and licensed?
Decision
Supreme Court ruled in favor of Google on fair use grounds.
Significance for AI Licensing
AI-generated software trained on existing code may:
Reproduce protected structures
Trigger fair use vs infringement debates
Influences licensing risk allocation clauses.
Case 6: SAS Institute v. World Programming Ltd. (EU, 2012)
Facts
Defendant developed software compatible with SAS system without copying source code.
Legal Issue
Are programming languages and software functionality copyrightable?
Decision
Court held that functionality and programming languages are not protected, only expression.
Significance
AI-generated software that replicates functionality may avoid infringement.
Licensing must focus on expression, not ideas or logic.
Case 7: GitHub Copilot Litigation (USA, Ongoing)
Facts
Plaintiffs alleged that AI-generated code reproduced copyrighted open-source code.
Claimed violation of open-source licenses.
Legal Issue
Does AI training and output infringe copyright and licensing obligations?
Current Position
Case ongoing, but courts are examining:
Attribution obligations
License compatibility
Derivative works doctrine
Significance
Directly impacts:
Licensing terms of AI-generated software
Compliance with open-source licenses
Risk allocation between AI providers and users
5. Key Legal Issues in Licensing AI-Generated Software
Ownership ambiguity
Who owns AI output: developer, user, or no one?
Authorship requirement
Most laws require human authorship
Patent inventorship
AI cannot be an inventor
Training data infringement
Risk of reproducing copyrighted code
License enforceability
Licenses may fail if no valid IP exists
6. Best Practices in Licensing AI-Generated Software
Clearly define human involvement
Specify ownership of outputs
Include indemnity clauses
Address open-source compliance
Allocate infringement risk explicitly
Use hybrid licensing models
7. Conclusion
IPR in licensing AI-generated software is evolving rapidly. Current legal position shows that:
AI cannot own IP
Human involvement is essential for enforceable rights
Licensing agreements must compensate for legal uncertainty
Courts prioritize human creativity, control, and accountability
Case laws demonstrate a clear trend:
👉 AI is a tool, not a legal author or inventor
Licensing strategies must therefore focus on contractual clarity rather than relying solely on statutory IP rights.
RELATED Blog
comments