Studio Ghibli and other Japanese publishers want OpenAI to stop training on their work

Look, Ghibli's right to be pissed, but they're six years late. OpenAI's been doing this since day one — ask forgiveness, not permission. That's now the industry standard.

Here's the thing: Studios got good press but zero leverage. They're not suing with teeth (yet), just complaining. Meanwhile, the model's already trained. The data's already in there. Deleting it doesn't un-train the weights.

What matters is whether they get a class action with real money behind it. Individual complaints are noise.