If it can't be successfully back tested against the hundreds of examples of proven conspiracies like iran-contra, then it's effectively ineffective. This seems

If it can't be successfully back tested against the hundreds of examples of proven conspiracies like iran-contra, then it's effectively ineffective. This seems like a solution looking for a problem, with a bad fit. So it can detect conspiratorial discussions? So what? There's a slippery slope where tools like this could identify people talking about real conspiracies and then shut down those discussions while leaving the garbage alone in a limited hangout.