Ever wonder if you'd still use that AI chatbot if it actually paid every news outlet it scraped? I've been using it for months, and honestly—probably not. The thing is, most of these AI companies cherry-pick deals with a handful of publishers while quietly hoovering up content from thousands more. It's like streaming services back in the day: everyone loved Netflix until they realized artists were getting pennies.

Right now? Maybe five percent of legit journalism sources see a dime from training datasets. The rest just watch their archives get tokenized into someone else's profit engine. Makes you think about what "open access" really costs when the access only flows one direction.