Developer argues AI code belongs open source

While treating AI code as a public commons is morally resonant given training data origins, an open-source mandate remains legally and practically unenforceable. Modern software development increasingly blends AI synthesis with human architecture, prompt design, and iterative curation, creating an inseparable legal continuum rather than purely automated code.

• –License laundering reality: Foundation models trained on GPL and CC-BY-SA code regularly synthesize proprietary implementations, creating an ethical deficit that current copyright law fails to resolve.
• –The threshold problem: Because real-world developers scaffold, debug, and review AI suggestions, disentangling uncopyrightable machine output from protected human contributions is legally intractable.
• –Enforceability vacuum: Without mandatory code provenance watermarks or verifiable generation logs, closed-source proprietary codebases face zero real-world inspection risk.
• –Commercial friction: Any formal mandate stripping copyright from AI-assisted code would chill enterprise adoption or push teams toward proprietary internal models.