> Yes, the "SA" (share alike) part means if you want to use and modify the words then you need to release them under license that is at least as permissive, but LLMs aren't using the words; they are clearly digesting the facts and expressing them in their own words

The CC BY-SA text <https://creativecommons.org/licenses/by-sa/4.0/legalcode.en> says nothing about the words. The words "word" and "words" do not even appear in the legal text. What it says, however, is:

> In addition to the conditions in Section 3(a) , if You Share Adapted Material You produce, the following conditions also apply.

> 1. The Adapter’s License You apply must be a Creative Commons license with the same License Elements, this version or later, or a BY-SA Compatible License.

> 2. You must include the text of, or the URI or hyperlink to, the Adapter's License You apply. You may satisfy this condition in any reasonable manner based on the medium, means, and context in which You Share Adapted Material.

where "Adapted Material" is defined as:

> material subject to Copyright and Similar Rights that is derived from or based upon the Licensed Material and in which the Licensed Material is translated, altered, arranged, transformed, or otherwise modified in a manner requiring permission under the Copyright and Similar Rights held by the Licensor. For purposes of this Public License, where the Licensed Material is a musical work, performance, or sound recording, Adapted Material is always produced where the Licensed Material is synched in timed relation with a moving image.

So if you want to argue that LLMs don't have to follow the SA clause, then you have to argue either that:

1. LLMs aren't an arrangement or transformation of their input, or

2. the license doesn't apply at all, eg. because it's fair use in whatever jurisdiction they fall under.

OpenAI is using argument 2 <https://news.ycombinator.com/item?id=37780199>.