1. Conceptual Foundation

“Algorithmically generated moral codes” refers to ethical or normative decision-making frameworks produced by computational systems—such as AI models, machine learning classifiers, or rule-based algorithms. These systems may generate:

• Content moderation rules (what is “harmful” or “acceptable”)
• Automated sentencing or bail risk scores
• Ethical decision trees in autonomous vehicles
• Recommendation systems embedding value judgments
• AI-generated “guidelines” for behavior or compliance

The legal question is:

Can such algorithmically generated moral frameworks be protected as intellectual property (IP) or treated as protectable intellectual works?

They sit at the intersection of:

• Copyright law (original expression)
• Trade secrets (confidential algorithms)
• Patent law (technical processes)
• Public policy (ethical governance and accountability)

Courts globally have not directly recognized “moral codes” generated by algorithms as a standalone IP category, but relevant principles emerge from software, AI authorship, and computational output cases.

2. Key Case Laws (Detailed Analysis)

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

Issue

Whether a simple compilation of data (telephone directory) can be protected by copyright.

Principle Established

• Copyright requires “minimal creativity”
• Mere “sweat of the brow” (effort) is not enough
• Facts and systems are not protected—only original expression is

Relevance to Algorithmic Moral Codes

Algorithmically generated moral frameworks often consist of:

• Rules derived from data
• Statistical weighting of ethical outcomes
• Automated classifications (e.g., “toxic content”)

👉 Under Feist, such outputs are likely not protected unless there is human creative input in selecting or structuring them.

Impact

If an AI generates a moral decision matrix automatically:

• It is likely treated as functional data/system, not copyrightable expression.

Case 2: Computer Associates v. Altai (1992, USA)

Issue

Whether non-literal elements of software (structure, sequence, organization) can be copyrighted.

Principle

Introduced the Abstraction-Filtration-Comparison Test:

1. Abstract structure of software
2. Filter out unprotectable elements (ideas, functionality)
3. Compare remaining expressive elements

Relevance

Algorithmic moral codes are typically:

• Structured decision trees
• Logic-based outputs
• Embedded policy rules

👉 Courts would likely:

• Strip away functional ethical rules (“if harm > threshold → block content”)
• Protect only expressive human-designed structure, if any

Key Insight

Purely functional ethical logic = not protectable expression

Case 3: SAS Institute Inc. v. World Programming Ltd. (2010, UK/EU influence)

Issue

Whether functionality of a software system and programming language behavior can be copyrighted.

Holding

• Copyright does NOT protect:
  • Programming languages
  • Data file formats
  • Functionality of software
• Only source code expression is protected

Relevance

If an AI system generates moral codes (e.g., moderation policies or scoring rules):

• The underlying ethical functionality is not protected
• Only literal code (if original) may be protected

Impact

This strongly limits IP protection over:

• Algorithmic ethical systems
• Automated governance rules

Case 4: Naruto v. Slater (Monkey Selfie Case, 2018, USA)

Issue

Whether a non-human can own copyright.

Holding

• Only humans can be authors under US copyright law
• Animal-created works are not protected

Relevance to AI Moral Codes

If AI generates moral frameworks autonomously:

• No “author” in legal sense
• Therefore, no copyright ownership unless:
  • A human can be identified as the creative controller

Key Principle

Non-human generated content lacks copyright ownership.

Impact

Fully autonomous moral algorithms = public domain by default

Case 5: Thaler v. Perlmutter (2023, USA)

Issue

Whether AI-generated artwork can be copyrighted without human authorship.

Holding

• Copyright Office and courts confirmed:
  • Human authorship is mandatory
  • AI alone cannot be an author

Relevance

Algorithmically generated moral codes (e.g., AI ethics models producing rules):

• Not protectable unless:
  • A human meaningfully designed or selected outputs

Important Distinction

• AI-assisted moral code → possibly protected
• AI-autonomous moral code → not protected

Case 6: Nova Productions v. Mazooma Games (2007, UK)

Issue

Whether frames generated by a computer program in a game could be copyrighted.

Holding

• The computer is merely an “extension of human skill”
• Copyright belongs to the human creator of the system
• Individual outputs generated by software are not separately authored works

Relevance

If an AI generates evolving ethical rules:

• The system owner may claim ownership only if:
  • They designed the system
  • Outputs are foreseeable results of human input

Key Principle

AI output is legally treated as:

“mechanical extension of human creativity”

Case 7: Google LLC v. Oracle America (2021, USA Supreme Court)

Issue

Whether copying software APIs constitutes fair use.

Holding

• Google’s use of Java API code was fair use
• Functional interfaces are less protectable than expressive code

Relevance

Algorithmic moral codes often function like APIs:

• “If X, then ethical action Y”
• Structured decision interfaces

Impact

• Ethical rule systems functioning as interfaces are likely:
  • Weakly protected or unprotected
  • Especially if they serve functional interoperability

3. Legal Synthesis: Protection of Algorithmic Moral Codes

Based on these cases, the legal position can be summarized:

A. Copyright Protection

Only possible if:

• Human authorship exists
• There is creative selection or arrangement
• Output is not purely functional or data-driven

Not protected if:

• Fully AI-generated
• Pure ethical logic systems
• Automated decision trees without human creativity

B. Patent Protection (Possible but limited)

Algorithmic moral codes may be patentable only if:

• They produce a technical effect
• Solve a technical problem (not abstract ethics)
• Are implemented in a novel system (e.g., autonomous vehicle safety system)

However:

• Abstract ethical reasoning is NOT patentable in most jurisdictions

C. Trade Secret Protection (Most realistic protection)

Companies often protect:

• Content moderation algorithms
• Ethical ranking systems
• Bias adjustment models

Requirements:

• Secrecy maintained
• Economic value derived from secrecy
• Reasonable protection measures

👉 This is currently the strongest protection route for algorithmic moral frameworks.

D. Public Policy Limitations

Courts are cautious because:

• Moral codes affect rights and liberties
• Transparency is required in governance systems
• AI ethics decisions may require accountability

Therefore:

• Over-protection is discouraged
• Open scrutiny often preferred in regulatory systems

4. Final Legal Conclusion

Algorithmically generated moral codes currently:

❌ Not independently protected as a distinct IP category

⚖️ Partially protected only through:

• Human authorship (copyright)
• Functional invention (patents)
• Confidentiality (trade secrets)

📌 Core Judicial Trend:

Courts consistently reject protection of:

• Pure functionality
• Machine-generated outputs
  • Abstract systems of rules without creative human authorship