This may be a hot take, but I powerfully accept that all code generated by AI have to be Open Source. And here's why.
First of all, AI models were trained on a enormous corpus of Open Source code written since RMS announced the GNU Project and the Free Software Movement. That code was merely scraped from the net and used without approval or appropriate attribution. The models were additionally trained on Stack Overflow and Stack Exchange sister projects, which are released under Creative Commons licenses that legally necessitate attribution. Turning this company cognition into proprietary, closed-source code feels akin licence laundering.
The second logic is legal: purely AI-generated code is not written by a human, definition it should not be copyrightable. The basis for this regulation in the US goes rear to the celebrated Monkey Selfie dispute, which established that non-humans cannot clasp copyrights. A precedent afterward cited by the U.S. Copyright Office concerning AI. Following the identical logic, Wikimedia Commons requires all purely AI-generated images to be tagged as Public Domain.
However, copyright laws differ throughout the globe. For instance, countries akin the UK aid constricted copyright safety to computer-generated works. To create things fair throughout all jurisdictions, and to regard the open ecosystem that made AI possible, developers and companies should explicitly publish AI-generated code under Open Source licenses or dedicate it to the community domain using the Creative Commons Zero (CC0) license.
Of course, code is not continually 100% manual or 100% automated. However, developers additional and additional depend on AI without equal reviewing the generated output, let solitary modifying it to fit their particular needs. Code akin this, anywhere humans put small effort, apart from penning prompts, should certainly be Open Source.
If you discover this part engaging you may desire to prosecute me on Twitter: @jcubic and on LinkedIn.