Because it isn’t in compliance with the terms of many popular licenses to create derivative works without attribution. Therefore the only possible claim Microsoft have to do so is via terms and conditions, which would not apply to code hosted on a stand-alone website.

It might be time for popular licenses to update themselves for this new-found threat.

> Therefore the only possible claim Microsoft have to do so is via terms and conditions

No, Microsoft's claim is fair use, not GitHub terms and conditions. You don't have to speculate on what you think their possible claims are, when they've publicly stated their actual claims.

What happens when some random person uploads your code (acquired from your website) to GitHub?

It depends on how that code is licensed. Possibly the random person gets served with a summons. Possibly nothing.

Well, that is what updating the license accomplishes.

I’d be interested to know if code subject to DCMA takedown would be removed from the copilot training set…

> it isn’t in compliance with the terms of many popular licenses to create derivative works without attribution

Did you click the link for this article? Did you see that MS removed the author’s name in the license header and replaced it with their own?

Yes?

It's a question of handing it over and giving them a ToS shield to hide behind, versus making them work for it and risking license violations "in the wild".

Does their ToS give them additional rights to the code uploaded to GitHub? There are several unofficial copies of projects like glibc [1] uploaded by people who definitely do not have the authority to grant any additional rights to the code.

[1]: https://github.com/bminor/glibc

Microsoft owns GitHub and you accept their terms of service.

Copilot is trained on public GitHub repositories of any license: https://en.m.wikipedia.org/wiki/GitHub_Copilot#Technology

If they scrape your website, that's different.

If it's legally sound to do that, then it's legally sound to scrape code from Stack overflow as well. No special license is granted to GitHub to train CoPilot; the hosting license in the ToS[0] specifically doesn't allow its use outside of GitHub itself[1], so i'd argue that applies to running copilot in VSCode for code not destined for GitHub - and i'm sure MS's lawyers reviewed such a product launch.

0: https://docs.github.com/en/github/site-policy/github-terms-o...

1: > This license does not grant GitHub the right to .. otherwise distribute or use Your Content outside of our provision of the Service