[philholden]

US copyright office seems to say anything written by AI cannot be copyrighted. Also a work that does not disclose which parts are AI generated cannot be copyrighted. All open source licences and creative commons are based on copyright. If you do not own the copyright you cannot release code under MIT.

As far as I can tell no one is labeling which parts of code used AI. Most devs I know are using Chat GPT at least for inspiration in solving problems. Pretty sure soon almost every file will have snippets of AI code. Companies have little control to stop this. So to me it looks like every company is unknowingly making its codebase public domain.

Open AI caches its responses so can scan private codebases to know which use AI. Copilot also knows which codebases use AI. If they find a codebase that uses AI but does not disclose where it is used then that codebase can be used for further AI training. Major companies will need to use copilot x, vscode and github to remain competitive. So microsoft could end up sucking up all the proprietary knowledge of every industry.

[thepostman0]

Competitive at what? Templating?