OwnershIP Of AI-Generated Predictive Climate Change Impact Simulation Models.
1. What is Being “Owned”?
AI-generated climate impact simulation models can include:
- Algorithms predicting sea-level rise, heatwaves, or flooding
- Scenario simulations for policy or infrastructure planning
- Data visualization models for decision-making
- Recommendations for mitigation and adaptation strategies
Potential IP protections:
- Patent Law – For novel methods, processes, or systems
- Copyright Law – For expressive components like code, dashboards, and reports
- Trade Secrets – For confidential predictive models
The main legal question is: Can AI be recognized as an inventor/author, or does ownership belong to humans or institutions?
2. Legal Principles
Human Authorship Requirement
- Most jurisdictions require a natural person as the inventor or author.
Work-for-Hire Doctrine
- If AI outputs are generated under employment, the employer may own them.
Inventorship vs Ownership
- Even if AI generates the model autonomously, ownership usually vests in the person/entity controlling the AI.
3. Key Case Laws
1. Thaler v. Commissioner of Patents
Facts:
Stephen Thaler’s AI, DABUS, autonomously generated inventions, and he requested AI be named as the inventor.
Issue:
Can AI legally be an inventor?
Judgment:
- Initially, the Australian Federal Court recognized AI as an inventor.
- On appeal, the court ruled: only humans can be inventors.
Relevance to Climate Models:
- Autonomous climate simulation models cannot list AI as inventor.
- Ownership defaults to developer or controller of the AI.
2. Thaler v. Vidal
Facts:
Thaler attempted to patent a DABUS-generated invention in the US.
Judgment:
- Courts rejected AI inventorship; only natural persons can be inventors.
Application:
- AI-generated climate simulation models cannot be patented in the AI’s name.
- Ownership resides with the human or corporate entity managing the AI.
3. Naruto v. Slater
Facts:
A monkey took a selfie; the question was about copyright ownership.
Judgment:
- Non-human entities cannot hold copyright.
Application:
- By analogy, AI cannot own copyright in climate simulation outputs.
- Human authorship or corporate ownership is required.
4. Feist Publications v. Rural Telephone Service
Facts:
Phone directory listings lacked originality and were not copyrightable.
Principle:
- Copyright requires minimal human creativity.
Application:
- If AI autonomously generates climate models without meaningful human input → may not qualify for copyright.
- If humans contribute creativity (e.g., model design choices) → copyright may apply.
5. Diamond v. Diehr
Facts:
Software controlling industrial processes was challenged for patentability.
Judgment:
- Software combined with a physical process is patentable.
Application:
- Climate simulation models integrated with real-world environmental interventions (e.g., adaptive infrastructure planning) may be patentable.
6. Mayo Collaborative Services v. Prometheus Laboratories
Principle:
- Laws of nature or abstract phenomena cannot be patented unless combined with novel applications.
Application:
- AI models that only simulate natural climate laws → not patentable.
- Models that provide inventive adaptation strategies → stronger patent eligibility.
7. Alice Corp. v. CLS Bank International
Principle:
- Abstract ideas implemented on computers are not patentable.
Application:
- Climate models purely performing calculations → abstract, less likely patentable.
- Models tied to practical climate adaptation solutions → potentially patentable.
4. Ownership Scenarios
- AI developed by a private company → Company owns outputs.
- Public research institution → Institution may own outputs; IP policies could allow open access.
- Human-AI collaboration → Humans contributing significantly may share ownership.
- Fully autonomous AI with no human input → Ownership defaults to AI owner/controller.
5. Special Considerations
(a) Data Ownership
- Climate AI relies on environmental, geographic, and demographic datasets.
- Data agreements or privacy regulations can affect ownership rights.
(b) Liability
- Even if an organization owns the simulation model, liability may fall on:
- Developers
- Policymakers using the predictions
- Implementing agencies
(c) Ethical Considerations
- Since models influence climate adaptation policy, transparency, reproducibility, and human oversight are essential.
6. Key Takeaways
- AI cannot legally own predictive climate simulation models.
- Ownership typically resides with humans or entities controlling the AI.
- Patent law is usually the most viable protection for AI-generated models with practical application.
- Copyright may only apply when there is human creative input.
- Courts consistently emphasize:
- Human authorship
- Practical, non-abstract applications
- Novelty and inventiveness
RELATED Blog
comments