[muzani]

I think it's a gray area in the license. Much of the code was intended to be used freely and commercially by others, but not for AI training. It follows the license to the letter, but not the intent.

I expect we'll see new licenses appear making it clear whether or not the content can be used for training.

[elil17]

Who's to say the intent? I've published lots of code with very permissive licenses and I did so because I want people to be able to use that code for any reason. That's why I choose those licenses.

[muzani]

I think that's exactly why AI training (allowed or not) should be added to licenses.

[josephcsible]

There's nothing gray about it. The license requires attribution, and Copilot doesn't provide that attribution.

[muzani]

It's reading the code and generating similar code, not copying it.

[josephcsible]

Are you saying that's fair use? If so, then we won't see new licenses appear related to it, since a license can only give you more permissions on top of fair use, not take away fair use. If not, then we still won't see new licenses appear related to it, since the existing licenses already don't allow it.

[muzani]

Good point. I'm not a lawyer, but looking it up, the factors for fair use are:

1. the purpose and character of the use; 2. the nature of the copyrighted work; 3. the amount and substantiality of the portion used; 4. the effect of the use upon the potential market for the original work.

All of these are quite debatable, and I'll leave it to someone more familiar with the law.

Though if it's not, I believe there are licenses that allow derivative uses of code and licenses that don't. For many of these, the intention is that they create more code, but not be used to fuel AI behemoths.