There are some promising ideas out there based on unsupervised ("self-supervised") learning. There, the problem of needing big labeled datasets doesn't exist: j

There are some promising ideas out there based on unsupervised ("self-supervised") learning. There, the problem of needing big labeled datasets doesn't exist: just turn on a camera and have a motor point it at the surrounding world. Data of this kind is and always has been super abundant. But using it requires making a conceptual leap away from supervised pattern recognition -- which is pretty much what today's "AI" does...