98
Tulsi Gabbard Fed Potentially Classified Documents Into AI to Determine If They Were Classified
(www.latintimes.com)
this post was submitted on 12 Jun 2025
98 points (98.0% liked)
Trump Watch
72 readers
474 users here now
Fascism has come to America, wrapped in a flag and carrying a cross.
Documenting the crimes and corruption of the 47th president of the United States and his fascist minions.
Here we go again. This man is a traitor.
Share articles related to Trump and MAGA here.
Liberalism, noun - “A political theory founded on the natural goodness of humans and the autonomy of the individual and favoring civil and political liberties, government by law with the consent of the governed, and protection from arbitrary authority.”
founded 1 week ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Just another show of incompetence. Like, that's not how AI works. I know people think that AI knows everything but it simply doesn't. And if you ask it a novel question that it has no answer for it doesn't just say "I don't know," it makes up random bullshit to fulfill its instructions as an LLM to answer any question. And there's literally no way for an LLM to know what's classified and what's not because it doesn't have the training data because it's, you know, classified. It's like a teacher asking a kid to write an essay on a book that they never read or was even instructed to read but saying that they can't just say "I never read that." You know what they'd turn in? Bullshit. And the fact that she fed classified information into an AI model is so incredibly stupid it's laughable because now that model has the classified information as training data and literally anyone can retrieve that information if they try hard enough.
We're rapidly devolving into a place where dumb people think AI is all-knowing, and you won't be able to convince them otherwise. Add in the fact that oligarchs build and control the LLMs, and you have a recipe for widespread manipulation of the populace that makes old school propaganda look quaint.
That is a great point to show how bad AI is, it won't say "I do not know" or "I cannot be certain that..."
And this, again, stems from the fact that AI has no idea about what it's doing, it had no idea about anything, it's just trying to find the shortest path to an answer and we've set it up in such a way that the shortest way statistically many times is the right way.