Fair Source Software in the AI age

The software licensing landscape faces unprecedented challenges from generative AI. Recent advances in LLM capabilities—particularly the “December Leap” of 2025—have created autonomous coding agents that fundamentally alter how software gets created and distributed. This shift happened not gradually, but specifically this last December.

The AI Disruption

Training Without Consent

LLMs were trained on publicly available source code with minimal regard for licensing requirements. Whether a codebase uses permissive (MIT, BSD) or restrictive (GPL) licenses, LLMs are not complying with license requirements. The fundamental challenge: enforcement is difficult when LLMs generate derivative works rather than exact copies.

The Rewrite Problem

AI agents dramatically lower the cost of reimplementing existing software. The chardet library controversy exemplifies this tension—a maintainer created a clean-room reimplementation under MIT instead of the original LGPL, arguing the rewritten code wasn’t a derivative work. With AI assistance, such rewrites become economically viable for any developer.

Copyright Uncertainty

The U.S. Supreme Court recently affirmed that “a significant human element is necessary” for copyright protection. AI-generated code without substantial human involvement may lack copyright protection entirely, creating a gray zone for licensing obligations.

Why Fair Source Still Works

Despite these disruptions, Fair Source software remains viable because it combines two protections:

• Simple non-compete clauses prevent direct competitors from using your software
• The eventual Open Source transition ensures code becomes freely available after a defined period

This two-part structure addresses AI-era concerns better than pure Open Source (vulnerable to rewrites) or proprietary models (unable to benefit from community contributions).

Human creativity and intentional licensing choices remain legally significant, providing Fair Source advocates defensible ground even as AI reshapes software creation.