I was looking at my parametric einstein / spectre tiles and noticed that @cults3d appears to have defaulted to a no-ai license on everything, since I don't recall setting it and I see it there.
I think that makes sense as a default global setting.
I'm also now faced with a conundrum: I licensed that work CC0 inasmuch as copyright even could apply to it, which isn't clear to me. To me, CC0 implies not restricting AI training.
I think that calling LLMs generalized AIs is bogus, and as much hype as OpenAI has built on the pretense, when things went bad they were happy to say "well ackshually it's just random probabilities," I've also known for a long time that using statistics to find hidden patterns is pretty much what the entire field of statistics is for, and ML and LLMs are fundamentally interesting and sometimes unreasonably effective applications of statistics.
So I think I really should not try to restrict training any form of statistical inference on CC0-licensed content, and I'm not convinced that I want to opt out for other works that I have licensed under what I intended to be open source terms.
The open source definition has famously excluded restrictions on field of use. Even as an LLM-as-generalized-AI utter sceptic, I feel like banning LLM training on work that I've released as open source would be inconsistent with my open source values.
GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.
All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.