Not looking forward to Plagiarism As A Service.

What's the status right now for whether LLM outputs can be copyrighted by the same big companies that ignored copyright to train them?