What I learned by arguing with Google AI about the Rucki case

AI treats court orders as gospel which means critical nuance is lost unless you present enough challenging evidence.

From Google AI about the Rucki case

Note: above is a video of a 2013 local Fox broadcast about the Rucki case.

Artificial Intelligence (AI) is on everyone’s mind; I decided to apply it to the story I have covered the most: the notorious Minnesota child custody case- the Rucki case.

The official version is that Sandra “Sam” Grazzini-Rucki was a vindictive woman who brainwashed her kids until her two oldest daughters ran and remained hidden for approximately two and a half years- April 2013 to October 2015.

All allegations of abuse against David Rucki were unfounded and the product of severe “parental alienation”, according to this version.

The official version is not only wrong but a deliberate lie by courts, police, and media like ABC’s 20/20.

From a child protective services (CPS) report in which Nico Rucki accused his father of sticking a gun to his head when he was eight years old.

AI is supposed to be really, really smart, and it should cut through the noise and get to the heart of the matter.

As such, examining the Rucki case is a perfect test case for the power of AI.

I started by asking a basic question: Is there any evidence that David Rucki abused his kids?

Google AI’s answer read like a press release from the courts or the police. It took the party line that those entities have conjured up for over a decade.

It even said, “Sandra successfully filed a two-year protective order, which he did not legally object to at the time, though no proof of physical abuse was submitted.”

This is legally impossible. If she successfully got a protective order granted, the court heard enough evidence of abuse to grant it.

I wasn’t someone with a passing interest in this case. I knew plenty, and I was ready to challenge AI.

I set out to prove AI wrong, and I did.