Too late, they are probably part of datasets for LLM training already as e.g. Google and Microsoft/OpenAI seem to use web content for training their models. But even if they're not used for training it won't help as you can just feed the articles to the LLM as context and ask it e.g. to write a similar article, ChatGPT can e.g. already fetch websites and work with the resulting text.

So unless we can get some better copyright protection laws enacted it seems that most content will be fair game for AI algorithms. Not sure I dislike that to be honest, it kind of moves the intellectual property frontier away from content and up into the idea level, i.e. text is not special anymore, it's rather the insights that produce the text that you want to protect. Interesting times we live in.