Hacker News Reader

What if they OCR’d the newspapers? No ToS there.

I’m pretty sure there is still a copyright also for the physical newspaper.

For the paper or the author? What exactly was the licensing agreement for Op-Ed authors in 1962?

Read the article. It's not difficult to get ChatGPT to regurgitate recent, obviously copyrighted articles, verbatim.

It will be equally easy for ChatGPT to rewrite copyrighted content that makes the output materially different for a copyright claim to succeed also.

Then ChatGPT should do that.

It's at least partially a copyright claim, isn't it? So the method -- OCR or scraping -- doesn't matter, I think.