using human-made works where the original authors didn’t consent to or even know that their work is being used to train the AI. Are there any initiatives to address this issue?

I don’t think so.

Training is expensive, and they all want to make money. Most of them want to make insane amounts of money.

Therefore they don’t care if it’s illegal.

Free and Nonprofit usually don’t start with huge piles of play money. OpenAI was the exception once before it got so thoroughly corrupted.