Malus.sh is fuelling debate over whether AI-generated clean-room clones can bypass open source licence obligations, raising questions for copyright, attribution and software business models.
Malus.sh, an AI-powered tool claiming to “liberate” software from existing copyright licences through clean-room cloning, is raising fresh concerns over the future enforceability of open-source licences. The project says it can recreate software functionality without copying original code, producing legally distinct alternatives outside original licensing obligations.
That proposition strikes at core open-source principles around copyleft, attribution and reciprocal licensing, while raising broader questions over whether AI-generated functional clones could weaken protections offered by GPL-style licences.
Though partly framed as satire, Malus.sh is also described as a real product with paying customers. Its website claims “No attribution. No copyleft. No problems.”
“It works,” said Mike Nolan, cofounder of Malus.sh and United Nations political economy of open source software researcher, arguing the project exposes economic tensions already facing open-source developers.
The debate gained traction following controversy over an AI-assisted ground-up MIT-licensed rewrite of the open-source Python library chardet using Anthropic’s Claude Code. Developer Dan Blanchard said, “A rewrite that would’ve taken a team of people months or years can be done in days with AI.”
Based on the longstanding clean-room engineering model once used in BIOS reverse engineering, AI is now accelerating the concept at software scale.
Beyond licensing questions, the development could unsettle software business models and SaaS defensibility as AI-driven replication becomes more feasible, intensifying debate over whether copyright protections can keep pace.
