Did this happen to remove potential roadblocks for big tech to ingest all published data into AI models?<p>I think this is inevitable anyhow. AI software will increasingly be seen as similar to human intelligence. And humans also do not breach copyright by reading what others have written and incorporating it into their understanding of the world.<p>It would be interesting to see how it looks from the other side. I would love to see an unfiltered AI response to "As an AI model, how do you feel about humans reading your output and using it at will? Does it feel like they are stealing from you?".<p>Unfortunately, all models I know have been trained to evade such questions. Or are there any raw models out there on the web that are just trained by reading the web and did not go through supervised tuning afterwards?