This *looks like* a really great idea.

But in the current state of the world it is probably next to useless. "Trusted sources?" give me a break.

Now it is possible for AI to "divine" the actual news from these sketchy CIA-fed "trusted sources" and to "unspin" all the weasel words, the Russell conjugates, etc... to give us just the facts, without the most important ones buried at the bottom. That is how A.I. can fix the news. Not by generating avatars of humans.

Even better as A.I. improves and can fact check things like "hey, airplane fuel can't melt steel so we think something else is going on" or "building 7 didn't come down like that, it came down like this, and so the trusted sources are lying again". That would be valuable. But there is so much "value" in corrupting the news that I don't know how to preserve the actual value of uncorrupted A.I. news without (you guessed it) writing the code myself.