AI-assisted fact-checking
Researchers developing AI to detect possible fake news and other manipulated content like doctored videos
Multiple types in development; one model uses info on how fake news spreads (more shares than likes, etc.), compares that to what it sees, then compares w/ other criteria (key words, etc.)
Traces fake news back to its roots to find more
Compiles data related to what it finds
Gives everything it finds (articles, source origins, data about topic) to journalists and fact-checkers to help them do their job more efficiently
True that tech isn't always reliable
AI would only identify possible sources of fake news; humans would manually verify