Why does nobody talk about who owns the training data?
Last Tuesday I sat through a 2 hour panel in San Jose where one guy kept saying AI art is just like a human learning from other art, and a woman in the back cut him off and asked who paid for the last 20 years of photos and forum posts those models ate up. Half the room nodded, half the room rolled their eyes, and I've been stuck on it since. If a model can only be as good as what it was fed, who should get paid for that food, and would you still use the tool if the answer was ordinary people like us?