No, not sue me for lemmy comments. AI is trained with lots of data. The world wide web is full of publicly accessible data like our comments. However, not all publicly accessible data may be used without a license. Examples thereof are news paper articles, videos, still pictures, etc. Normally, if you want to use those commercially, consent has to be given by the license holder and a in some cases a fee has to be paid.
Microsoft Copilot is an AI model to help people write code. However, it was trained mostly on opensource code (code made publicly available) which was very often licensed. And it is done so in such a manner that commercial use is allowed with the obligation to make that commercial code publicly available too. Microsoft does not make the code for Copilot publicly accessible and uses code licensed in many, many other ways - and it does so without asking for consent.
This is often a double standard as companies that hide their code fight very hard to keep it secret and/or pursue those in court who do not get a license to use it. However, they will happily use licensed consent to their benefit without consent nor potential payment.
With some clever tricks, AIs have been duped into revealing their training data (often licensed, sometimes very private e.g addresses, birthday, health information, etc.). Lawsuits have ensued (against the AI owners like Microsoft) and are currently active with a pending verdict. Until the verdicts come, I add the license link to my comments. Who knows, maybe it will have an impact, maybe not.
Hopefully I could explain the situation in an understandable manner for you.
The worst documentation of a linux distro I have ever encountered, but the declarative model has convinced me I don’t want something else. Now I’m just waiting for other distros to pop up that are declarative as well. (Guix? No thanks, I’m not a fan of endless parentheses)