AI And The Ship of Theseus

2026-03-05 · Source: Armin Ronacher's Thoughts and Writings · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Emerging Technologies & Innovation · Depth: Intermediate, short

Summary

Armin Ronacher's March 05, 2026, article explores the implications of increasingly cheap code generation, particularly through AI and re-implementations from test suites, on software licensing and copyright. He cites an instance where an AI ported his library to another language, resulting in a functionally similar but structurally different design. A similar situation occurred with the `chardet` library, which was re-implemented from scratch using only its API and test suite to enable relicensing from LGPL to MIT. This re-implementation, despite being distinct by JPlag validation and offering significant performance improvements, sparked a debate over whether it constitutes a derived work. Ronacher highlights that this trend challenges copyleft licenses like GPL, as code can be trivially rewritten, potentially leading to both open-source and proprietary abandonware re-emerging under new licenses. He also raises the possibility that AI-generated code might not be copyrightable due to insufficient human input.

Key takeaway

For CTOs and VPs of Engineering navigating software licensing and open-source strategy, the decreasing cost of code generation and AI-driven re-implementations necessitate a re-evaluation of intellectual property protection. You should assess your reliance on copyleft licenses and consider the potential for "slopforks" to emerge, challenging existing copyright claims. Explore strengthening trademark protections for your projects as a more resilient alternative to license enforcement.

Key insights

Cheap code generation and AI challenge traditional software licensing and copyright, enabling re-implementations from test suites.

Principles

• Copyright enforcement relies on friction.
• Trademarks may be more robust than licenses.

Method

Re-implementing software from its API and test suite can create a functionally identical but structurally distinct work.

In practice

• Consider re-implementing GPL libraries for permissive licenses.
• Evaluate trademark protection for software projects.

Topics

• AI Code Generation
• Software Licensing
• Copyright Law
• Open-Source Licenses
• Code Re-implementation

Code references

• chardet/chardet

Best for: CTO, VP of Engineering/Data, Director of AI/ML, Software Engineer, AI Engineer, Legal Professional

Related on AIssential

• If the machine can regenerate it, the old social contract doesn’t matter. That’s the same philosophical move you see in other domains: ingestion without consent, then “the output is new.” — AI-driven "clean-room laundering" challenges traditional IP and open-source licensing by enabling cheap, deniable code replication.
• Disney + Sora: Billion Dollar Partnership Signals New Era — Major IP holders are shifting AI training from open access to high-value licensing deals, using litigation as leverage.
• Who Owns the Code Claude Wrote? — AI-generated code ownership is legally complex, unsettled across human authorship, employer IP, and open-source license contamination.
• Why Europe's latest Meta ruling matters: International approaches to AI legislation are diverging. — Europe and the US are diverging on AI legislation, particularly regarding platform compensation for content extraction.
• Of Slop and Swarms: The First Amendment's Next Test — First Amendment protection for AI outputs should hinge on demonstrable human expressive contribution and output predictability.
• Industry welcomes rethink as UK drops preferred AI copyright plan - Research Information — The UK government dropped its preferred AI copyright exception, prioritizing fair licensing and creator protection.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Armin Ronacher's Thoughts and Writings.