Newsrooms using AI to make some money off local archives? It's poisible, but ...

 This is interesting. Should newsrooms train chatbots on their own content in an attempt to capture traffic from answering locally oriented questions/research (and some $$ in the process)?

https://rjionline.org/news/building-a-chatbot-trained-on-your-newsrooms-content/

It's an intriguing concept, but I'd suggest that except for the largest legacy newsrooms, one's content is not enough to provide the comprehensiveness (breadth and historical depth) to avoid bias and minimize troubling artifacts. That's especially because we know how bad many "print" newsrooms have been at preserving content and digitizing older materials. Broadcast is even worse.

But a local cooperative where everyone can get a share of the proceeds might work, though we also know hard those have been to put together in recent years.

(And before you ask, I can see big chains trying to do it across their properties, but thats's not the point here, and I actually don't think it would work that well. Do I really want to search across Gannett? Or do I really want to go deep across South Succotash? I'd suggest it's the latter.)

This is where libraries could have a key role as nucleii oi such efforts. But, again, we know how difficult it is to get a buch of like-minded (as in make money) news orgs to cooperate. Throw a nonprofit/governmental org into the mix and it becomes exponentially complicated, with the chances of the for-profit orgs sloughing off stuff on the NP.