There does need to be a mechanism to keep the human in the loop to correct the knowledge base by people who have the underlying knowledge. Perhaps notification needs to be sent to people who have previously viewed the incorrect information when a correction is made.

Did you even try diffy.chat to test how factually correct it is and how well it cites its sources? How good does it have to be to be useful? How bad does it have to be to be useless?

The language model for diffy.chat has been trained not to respond from its own learned parameters, but to use the Diffbot external knowledge base. Each sentence or paragraph in a Diffy response has a link to the source of the information.

They succeeded in getting attention. Look at all the comments posted here. The issue needs attention. The issue also needs fact checking. I was pleased with fact checking I got from diffy.chat about the wildfires in LA County. Maybe fact checking bots should be included in online discussion forums.