For the last forty years, business logic was simple: "If my employee writes it, I own it." That assumption has fractured. Recent US Copyright Office guidance and federal rulings (e.g., Thaler v. Perlmutter) have clarified that works lacking meaningful human authorship (inclusive of AI generated content) are not eligible for copyright protection.
If your "secret sauce", the core matching engine or proprietary algorithm, was generated by GitHub Copilot or ChatGPT, it may effectively be functionally uncopyrightable, limiting your ability to assert exclusive rights. You may lack standing to sue a competitor for copying code that is not legally protectable. Furthermore, if an AI output substantially reproduces GPL-licensed code, your entire proprietary application could be "infected," forcing you to open-source your product. It will become much more common to see lawsuits of “copyright infringement” for code such as this case with GitHub Copilot, a coding assistant tool.
You must shift from "Default Ownership" to "Strategic Provenance." You need an audit trail that proves exactly where the machine stopped and the human began.
This is how your team will execute the email at the bottom of this newsletter. Depending on your line of business and products, your team will use some variation of this protocol.
To protect your valuation, implement a "Zone-Based" Development Protocol.
🔴 Red Zone (Human Only): Your core differentiator (the top 10% of your code that drives value).
Rule: No AI generation allowed. Human-written only to preserve a clean, defensible chain of title.
🟡 Yellow Zone (AI-Augmented): Boilerplate, UI scaffolding, testing scripts.
Rule: AI allowed only with Enterprise licenses (no data training). Developers must document substantial human authorship sufficient to support copyright claims.
🟢 Green Zone (Commodity): Open-source components, generic documentation.
Rule: AI allowed freely. No copyright expectation.
"But doesn't Microsoft indemnify us?" Not always. We analyzed the fine print of the major AI "Copyright Shields." Most contain exclusions for "modifications" or failure to use specific technical filters. If you customize the code (which you always do), the shield may vanish
We have created a decision framework to help you evaluate your risk in real-time.
The Asset: The AI Ownership & Risk Decision Matrix
Format: PDF Decision Guide.
Includes: The "Can I Copyright This?" Decision Tree, Vendor Indemnity Trap Checklist, and the "Red Zone" Definition Guide.
Does your team need help assessing AI ownership and risk? Reply to this email with "AI Copyright" and we will send you the PDF mentioned above.
We drafted the email below to help you ask the right questions to your team. Copy and paste the text below to your Decision Science team and CTO to initiate an immediate IP audit.
To: CTO; Decision Science Lead
From:
Subject: Audit of AI-Generated IP & Ownership Risk
Team,
I am reviewing our exposure regarding intellectual property ownership of AI-generated assets.
Given recent legal rulings that AI-generated content cannot be copyrighted, I want to know how we are handling the usage of AI to generate assets.
Action Required: Please reply to this email addressing these three questions:
1. Core IP Isolation: Can we demonstrate meaningful human authorship for our ‘Crown Jewel’ algorithms, or do we have logs showing substantial human modification of AI outputs?
2. Indemnity Status: Are we relying on "Copyright Shields" from vendors like Microsoft or OpenAI via our licenses? If so, have we technically enabled the specific filters (e.g., "Block Public Code") required to keep those indemnities valid?
3. AI Usage: How are we currently leveraging AI for coding and generating our most crucial assets? Who is overseeing this to make sure we are retaining our IP?
We need to ensure we actually own the product we are selling.
Best,
