IP Issues In AI-Created Ethical Compliance Scoring Frameworks For Corporations.

13 Mar 2026 --
0 Comments

1. Background: AI-Created Ethical Compliance Scoring Frameworks

Corporations increasingly use AI to assess and score ethical compliance across operations, supply chains, HR practices, and environmental impact. These AI frameworks often rely on:

Proprietary algorithms

Training data (corporate policies, regulatory texts, prior audit reports)

Scoring methodologies (weighted metrics, risk indices)

Output dashboards and reports

IP concerns arise because AI frameworks combine multiple protected elements: software, algorithms, data, and potentially copyrighted content. Key issues include:

Copyright – Are the AI models’ outputs protected as “original works”? Can the underlying training data infringe third-party copyright?

Patent – Are scoring algorithms patentable, especially if they involve technical AI methods?

Trade Secrets – Protecting the proprietary methodology and data sources.

Ownership – Who owns IP when AI is used to create scoring methodologies: the AI developer, the corporate client, or is it unprotectable under law?

2. Key Case Laws and Their Analysis

Case 1: Thaler v. Comptroller General of Patents (US, 2021)

Facts: Stephen Thaler argued that an AI system, “DABUS,” should be recognized as the inventor for patent purposes, including for AI-generated methodologies.

IP Issue: Ownership of inventions created by AI. Can AI itself be an inventor under patent law?

Outcome & Implications:

US Patent Office rejected the claim; the inventor must be a natural person.

Courts confirmed AI cannot hold patent rights.

Relevance: For AI-created ethical scoring frameworks, corporations cannot claim AI as the inventor; humans or the commissioning company hold rights.

Takeaway: Intellectual property protection must be assigned to the human operator or company, not the AI system itself.

Case 2: Authors Guild v. Google (US, 2015)

Facts: Google scanned millions of copyrighted books to create a search engine. Authors sued for copyright infringement.

IP Issue: Use of copyrighted content for AI training or algorithm development.

Outcome & Implications:

Court ruled Google’s use was “transformative” and constituted fair use.

Relevance: AI training for ethical compliance scoring may similarly use copyrighted corporate or regulatory data. If the use is transformative and not redistributive, courts may consider it fair use—but this is fact-specific.

Takeaway: Companies must evaluate whether training datasets include copyrighted content and whether AI-generated outputs constitute transformative use.

Case 3: SAS Institute Inc. v. World Programming Ltd (UK & EU, 2012)

Facts: WPL developed software compatible with SAS’s statistical programs without copying code. SAS claimed copyright infringement.

IP Issue: Protection of software functionality vs. copyright in code.

Outcome & Implications:

EU Court ruled functionality and algorithms are not copyrightable; only the expression (code) is.

Relevance: Ethical scoring algorithms are functional systems. Merely replicating their functionality (e.g., scoring logic) without copying code may not infringe copyright.

Takeaway: IP protection for AI frameworks is strongest in the code and unique presentation of scoring, less so in the abstract methodology.

Case 4: Trade Secret Case – Waymo v. Uber (US, 2018)

Facts: Waymo sued Uber for misappropriation of trade secrets related to self-driving algorithms.

IP Issue: AI and algorithmic processes as trade secrets.

Outcome & Implications:

Uber settled; trade secret misappropriation can occur if employees leak AI-trained algorithms.

Relevance: AI-generated compliance scoring frameworks can be protected as trade secrets if kept confidential and access-controlled. Sharing datasets or code externally without safeguards risks litigation.

Takeaway: Corporations must implement strict confidentiality agreements, data handling, and access policies to protect AI IP.

Case 5: Feist Publications v. Rural Telephone Service (US, 1991)

Facts: Feist compiled a telephone directory with factual data; Rural claimed copyright infringement.

IP Issue: Copyrightability of compilations of facts.

Outcome & Implications:

Court ruled facts themselves are not copyrightable; only original selection/arrangement is.

Relevance: AI ethical scoring frameworks often compile regulatory rules, corporate data, and metrics. The raw data may not be protected, but the selection, weighting, and reporting logic can be IP-protected.

Takeaway: AI frameworks must differentiate between raw data and original methodology for IP protection.

Case 6: Oracle v. Google (Java APIs, US, 2021)

Facts: Google used Java APIs in Android; Oracle claimed copyright infringement.

IP Issue: Copyright protection for interfaces and functional elements.

Outcome & Implications:

Supreme Court ruled Google’s use of APIs was fair use.

Relevance: For AI ethical scoring frameworks, using standardized regulatory metrics or compliance scoring metrics may be legally safer than copying proprietary algorithms verbatim.

Takeaway: IP protection should focus on proprietary enhancements, not standard industry methodologies.

3. Summary of Legal Principles for Corporations Using AI Scoring Frameworks

IP Concern	Legal Guidance
Copyright	Protect code, dashboards, report templates; be careful with training data.
Patent	Only patent if the AI algorithm solves a technical problem in a novel way; AI cannot be listed as inventor.
Trade Secret	Keep proprietary methodology, scoring weights, and data confidential.
Ownership	Clearly define ownership in contracts between AI developer and corporate client.
Liability	Misuse of third-party data may cause infringement claims; maintain licensing or use public/fair-use data.

Practical Recommendation:
Corporations should combine copyright, patent, and trade secret strategies, while ensuring licensing compliance for training datasets. Contracts should clearly assign IP ownership, and AI-generated outputs must have human oversight for enforceability.