[voiper1]

Surely there's AI usage that's not morally reprehensible.

Models that are trained only on public domain material. For value add usage, not simply marketing or gamification gimmicks...

[qingcharles]

How many models are only trained on legal[0] data? Adobe's Firefly model is one commercial model I can think of.

[0] I think the data can be licensed, and not just public domain; e.g. if the creators are suitably compensated for their data to be ingested

[Eisenstein]

> How many models are only trained on legal[0] data?

None, since 'legal' for AI training is not yet defined, but Olma is trained on the Dolma 3 dataset, which is

1. Common crawl

2. Github

3. Wikipedia, Wikibooks

4. Reddit (pre-2023)

5. Semantic Scholar

6. Project Gutenberg

* https://arxiv.org/pdf/2402.00159