The present considering is that the results of accepting verbatim the output of a easy immediate isn’t copyrightable, and that nobody truly owns the code — an attention-grabbing notion in and of itself.
However then the moral query comes into play. If I discover a bug in an open supply mission, I ask GitHub Copilot to repair it, and Copilot writes a intelligent and efficient repair, then who cares who owns the code? Ought to a maintainer of the mission reject such a pull request simply because it was AI-generated? That appears foolish to me, but it’s occurring immediately.
Our code
There may be, too, the difficulty of license compliance for AI-generated code. As a common rule, LLMs generate code moderately than copying it. They don’t copy and paste code immediately from repositories. Nonetheless, there have been circumstances the place AI-produced code has resembled open supply code so intently that the declare could possibly be made that it’s a copy. If this occurs with GPL code, it could possibly be a violation of the license to make use of it with out the receiving code base being “contaminated.” Open supply maintainers naturally ought to be involved about this occurring.
