> Grist for the AI mill
Two tweets from Dr Ruth Alison Clemens:
Tweet one · 17 Jul 2024:
Just found out that Taylor & Francis has sold access to all @routledgebooks data to Microsoft to train their AI. This includes my publications. I get no payment for using my research labour, never mind the bigger problems of this energy-intensive extractivism...
Tweet two · 27 Sep 2024:
UPDATE: yesterday @routledgebooks @tandfonline told all staff that it's extra important to meet 2024 targets as they have promised the LLM companies a quota of books for them to digest to train their AIs. So if your editor is pushing you to meet that deadline, this is why!
AI models ingesting technical and academic texts is much better than training the models on the non-sequiturs of random Internet yahoos who aren't even told (so long as those individual academics are aware, compensated, and given an opt-out). Setting quotas for academics to feed the AI seems like madness, turning difficult-to-produce text into listicles that are published just for the sake of publishing.
The likelihood of an ouroboros of using AI to generate text to meet an AI training quota seems ever more likely.