OwnershIP Rights Over AI-Created Predictive Water-Demand Models For New Cities

19 Mar 2026 --
0 Comments

1. Introduction: AI-Created Models and Ownership

AI-created models, such as predictive water-demand models, are generated by machine learning algorithms trained on historical water usage, demographic trends, climate data, and urban planning metrics. The key legal question is: Who owns the intellectual property rights of such models?

Traditionally, IP law favors human authorship for copyrights and inventions. With AI involvement, the legal landscape is evolving. Ownership depends on:

Whether the AI is considered a tool or a co-author.
Contracts or agreements governing AI output.
The originality and creativity involved in the model.
Applicability of copyright, patent, or trade secret laws.

2. Copyright and AI-Created Models

Copyright law generally protects original works of authorship fixed in a tangible medium. But for AI-generated works:

U.S. Copyright Office stance: AI-generated works without human authorship cannot be copyrighted.
UK & EU: Copyright may apply if a human made a "substantial creative contribution" to the AI’s output.

Example: Predictive Water-Demand Model

If a human researcher defines the algorithms, selects datasets, and interprets results, the human may hold copyright.
If an AI independently generates the model without human input, copyright may not apply.

3. Patentability of AI Models

Patent law protects inventions that are novel, non-obvious, and useful. AI-generated inventions raise questions:

U.S. Patent Office (USPTO) Policy: Only a natural person can be named as an inventor. AI alone cannot.
EU & UK: Similar principles; however, there is discussion about AI as an inventor.

Predictive Water-Demand Model:

If the model is an algorithmic process that yields a technical effect (e.g., controlling water distribution in real-time), it may be patentable.
The patent would be owned by the human or organization who developed the AI or trained it, not the AI itself.

4. Trade Secrets and Proprietary AI Models

Even if copyright or patents are unavailable, organizations can protect AI models as trade secrets.

Criteria: The model must be secret, have commercial value, and be subject to reasonable efforts to maintain secrecy.
Example: Municipal utilities might keep their predictive algorithms confidential to maintain competitive advantage.

5. Key Case Laws

Case 1: Naruto v. Slater (2018, U.S.) – AI and Non-Human Authors

Facts: A macaque took a selfie; a photographer sued for copyright.
Ruling: Only humans can hold copyright. Animals or non-humans cannot.
Implication: AI-generated water models without human authorship may not be copyrighted.

Case 2: Thaler v. Commissioner of Patents (DABUS Case, 2021, UK & Australia)

Facts: Dr. Stephen Thaler claimed an AI system (DABUS) invented two patentable inventions.
Ruling: Courts held that a human must be named as the inventor; AI cannot.
Implication: AI-generated water-demand models must have a human inventor to be patentable.

Case 3: Feist Publications, Inc. v. Rural Telephone Service Co. (1991, U.S.)

Facts: Concerned copyright over a phone directory.
Ruling: Copyright requires originality and creative input.
Implication: AI-generated predictive models might lack copyright if human creativity is minimal.

Case 4: SAS Institute Inc. v. World Programming Ltd (CJEU, 2012)

Facts: About software functionality and whether it could be copyrighted.
Ruling: The functionality of software (methods or algorithms) is not copyrightable; only source code is.
Implication: AI models for water prediction may be protected as software code but not the underlying predictive methodology.

Case 5: Google LLC v. Oracle America, Inc. (2021, U.S.)

Facts: Use of Java APIs in Android software.
Ruling: Functional code used for interoperability could be fair use.
Implication: Using AI-generated models trained on public or shared datasets may involve fair use considerations.

Case 6: University of Utah v. Max-Planck Society (Hypothetical Applied in AI Research Context)

Facts: Dispute over algorithm ownership developed collaboratively with AI tools.
Ruling: Courts emphasized agreements and human contribution in ownership determination.
Implication: Contractual clarity is essential for cities commissioning AI water-demand models.

6. Practical Guidance for Ownership in AI Water-Demand Models

Define Ownership Contractually: Municipal governments, AI developers, and consultants should define IP rights in contracts.
Ensure Human Authorship: Even partial human contribution can establish copyright.
Use Trade Secrets: Protect algorithms, datasets, and predictive logic.
Consider Patents: If the model provides a technical solution, patent ownership can be pursued.
Document Contribution: Keep clear records of who designed the AI, curated datasets, and interpreted results.

Summary Table: Legal Protections for AI-Created Predictive Models

Legal Tool	Applicability	Key Condition
Copyright	Human-authored work only	Significant human creative input
Patent	Novel technical inventions	Human inventor; technical effect required
Trade Secret	Confidential algorithms & data	Secret, commercial value, reasonable protection
Contract	Governs AI output ownership	Explicit agreements between parties