1. Understanding the Topic

What are AI-generated pest-migration simulations?

These systems use AI/ML to predict:

• Movement of agricultural pests (locusts, fall armyworm, aphids, etc.)
• Weather + climate-driven spread patterns
• Crop vulnerability zones
• Seasonal infestation forecasting
• Geographic simulation maps

They rely on:

• Satellite data
• Sensor data from farms
• Historical pest outbreaks
• Climate models
• Machine learning predictions

2. Core Legal Questions

(A) Can AI-generated simulations be copyrighted?

(B) Are predictive models “facts” or “expressions”?

(C) Who owns the output—developer, user, or AI system?

(D) Can raw agricultural data be protected?

(E) Are predictive outputs too functional for copyright?

3. Legal Classification Problem

AI pest simulations sit between:

4. Important Case Laws (Detailed Explanation)

1. Feist Publications v. Rural Telephone Service (1991, USA)

Principle: Facts are not copyrightable

• Telephone directory was copied
• Court ruled:
  • Facts = free for public use
  • No protection for “sweat of the brow”

Application to pest simulations:

• Pest migration data (e.g., “locusts moved north”) = facts
• Climate and crop data = non-copyrightable
• AI cannot own underlying agricultural reality

Legal takeaway:

Raw agricultural and ecological data used in simulations is not protected.

2. Baker v. Selden (1879, USA)

Principle: Idea–expression distinction

• Accounting system described in a book was copied
• Court ruled:
  • Ideas, systems, methods are NOT copyrightable
  • Only expression of explanation is protected

Application:

AI pest migration models:

• Mathematical prediction model = idea/system
• Simulation visualization = possibly protected expression

Key rule:

You cannot copyright the method of predicting pest migration—only its expressive presentation.

3. Lotus Development Corp. v. Borland (1996, USA)

Principle: Functional systems are not protected

• Software menu structure was copied
• Court ruled:
  • functional interface = not copyrightable

Application:

AI pest simulation dashboards:

• forecasting interfaces
• map controls
• alert systems
  → may be considered functional tools, not expressive works

4. Feist + Compilation Doctrine (Combined principle)

Principle: Selection/arrangement must be creative

• Even databases require creativity

Application:

If a pest simulation system:

• only automatically compiles data → not protected
• if scientists curate:
  • regions
  • thresholds
  • risk classifications
    → may be protected as a compilation

5. Google LLC v. Oracle America (2021, USA Supreme Court)

Principle: Functional use and fair use in software context

• Java API structure was copied
• Court ruled:
  • copying functional code for interoperability = fair use

Application:

AI agricultural simulation systems:

• may reuse existing prediction models if:
  • used for interoperability
  • not direct substitution

Key insight:

Functional reuse in scientific AI systems may be allowed under fair use doctrines.

6. CCH Canadian Ltd. v. Law Society of Upper Canada (2004, Canada)

Principle: Skill and judgment required for originality

• Legal materials were copied
• Court ruled:
  • originality requires intellectual effort

Application:

Pest migration simulations:

• human scientists selecting:
  • model parameters
  • weighting climate variables
  • risk thresholds
    → may create copyrightable structure

But:

• fully automated AI predictions → no human authorship

7. Feist + Database Rights (EU BHB v William Hill, 2004)

Principle: Investment in data collection matters (EU law)

• Horse racing database copied
• Court ruled:
  • protection exists for substantial investment in obtaining data

Application:

Agricultural AI systems:

• collecting satellite + farm sensor data:
  • may qualify for database protection in EU
• but:
  • AI-generated predictions themselves are not “obtained data”

8. Football Dataco Ltd v. Yahoo! UK Ltd (2012, CJEU)

Principle: no protection for purely mechanical compilation

• football fixtures list dispute
• Court ruled:
  • no creativity → no protection

Application:

If pest migration simulation:

• automatically generated from climate feeds
  → no copyright protection

If curated by experts:
→ possible protection of dataset structure

9. Thaler v. Perlmutter (2023–2024, USA AI authorship case)

Principle: AI cannot be an author

• AI-generated artwork denied copyright
• Court confirmed:
  • human authorship is required

Application:

AI pest migration simulations:

• fully automated predictive maps:
  → not copyrightable
• human-supervised modeling:
  → possibly protected

10. Nova Productions v. Mazooma Games (2007, UK)

Principle: computer-generated outputs need human authorship linkage

• game frames were generated by software
• court ruled:
  • software outputs are not authored by users automatically

Application:

AI pest prediction outputs:

• generated by system = not owned by end-user
• possible rights belong to:
  • software developers
  • or curating scientists

11. Sega Enterprises v. Accolade (1992, USA)

Principle: reverse engineering allowed for interoperability

• copying code for compatibility allowed

Application:

Agricultural AI:

• using existing pest models to improve interoperability:
  • may be lawful if transformative
• encourages innovation in climate-agriculture AI systems

5. Key Legal Characterization of AI Pest Simulations

A. Raw agricultural data

• ❌ Not copyrightable (Feist)

B. Mathematical models

• ❌ Not copyrightable (Baker v Selden)

C. Software code

• ✔ copyrightable

D. Simulation outputs (maps, graphs)

Depends:

E. Dataset compilation

• ✔ possible protection (EU database right)
• depends on investment and human selection

6. Ownership Scenarios

Scenario 1: Fully autonomous AI system

• No copyright owner
• Output = public domain-like status

Scenario 2: Scientist-curated model

• Scientist may own:
  • structure
  • visualization
  • annotations

Scenario 3: Company-built system

• Company may own:
  • software code
  • database structure
• Not necessarily predictions themselves

7. Special Issue: Functional Nature

Courts consistently treat predictive simulations as:

functional scientific tools rather than artistic expression

This reduces copyright protection significantly.

Key cases supporting this:

• Baker v Selden
• Lotus v Borland
• Football Dataco v Yahoo

8. Final Legal Conclusion

AI-generated pest-migration simulations are generally:

❌ NOT protected when:

• purely AI-generated
• based on factual agricultural data
• automated without human creativity

✔ protected when:

• humans design the model structure
• scientists curate outputs
• datasets involve substantial curated investment
• software code is original

9. Core Legal Principle Summary

The more “scientific and functional” the AI simulation becomes, the weaker the copyright protection; the more “human-curated and expressive” it becomes, the stronger the protection.