"Don't talk about the assassination attempt"

[ArMaP]

Yup, I commented without seeing that image.... who published it I wonder?

Apparently, it was first published on "Threads" (never heard of that) and "X".

https://www.threads.net/@canadianguy56/post/C9aRR4FP737
https://www.threads.net/@ben.rockson/post/C9aTU7aJVQN
https://x.com/daviddunn177/status/1812463663778783381

Who "decided" to submit that photo as a part of the historical record?  Was that Meta?

As far as I understand it, Meta had nothing to do with it, I don't even know if someone posted that image on Facebook or not.

And, without having access to them both, how does the "AI" know which is which?

Pattern finding and matching has been used for a long time and is one of the most common AI uses.

Someone has to curate the images, was it a human that said.. this (smiling image) is an image of the event?

Yes, the "fact checkers" are human, they flag the images as fake. AI looks at other images and tries to see if they are the same image that was flagged as fake. That's what failed in this case, the AI didn't consider smiling/non smiling as a difference and said the original photo was fake.

I'm sorry... but I still feel a disconnect between reality and what we are presuming to be a programming tool that allegedly reports facts and reality.

A programming tool? What are you talking about?