Easier said than done.

How to distinguish facts from fake with current AI?

Especially in the future when every source has AI content.