Everything tech giants will hate about the EU’s new AI rules | EU rules ask tech giants to publicly track how and when AI models go off the rails.

* EU explicitly bans using torrented/pirated materials for AI training – directly calling out Meta who argued individual books are “worthless” for training after getting caught using pirated book datasets
* Code pushes for letting websites block AI crawlers without affecting Google search rankings – addressing the current impossible choice between SEO and protecting content from AI scraping
* Companies must reveal total energy consumption for both training and inference
* Any “serious incidents” including deaths, mental health harms, or infrastructure disruptions must be reported to EU within days – treating AI failures like airplane crashes
* Companies must detail exactly where training data came from (public vs user data vs synthetic)

Fkin love it. Pretty much everything you’d do when you’re not in anyone’s pocket. We’ll see how it will work globally though, or the extent it will be enforcable.