When AI goes lib-crazy (update: USA judges used AI to make 'rulings' with fake, hallucinated citations)

blablarg18

Robby Starbuck asked Meta AI about himself & it made shit up.

https://x.com/robbystarbuck/status/1917276375297232908

Things he can prove are impossible, like his location on J6.

How these things work: AI trains on lib echo chamber. Lib echo chamber, as we know, amplifies made-up shit.

Elon says he'll have to reset Grok AI training, for that reason.

https://x.com/elonmusk/status/1936333964693885089

https://x.com/elonmusk/status/1936493967320953090

Grok trains on "entire world" and sadly, that includes massive made-up massive lies of lib echo chamber. Google & Meta would only be worse.

Let this warn u. Check if AI makes up shit on you.

blablarg18

Google AI is wild! & Starbuck sues

https://x.com/robbystarbuck/status/1980985160586625186

Since 2023, @GoogleAI
(Bard, Gemini & Gemma), has been defaming me with fake criminal allegations including sexual assault, child rape, abuse, fraud, stalking, drug charges, and even saying I was in Epstein’s flight logs.

All 100% fake. All generated by Google’s AI. I have ZERO criminal record or allegations.

So why did Google do it? Google’s AI says that I was targeted because of my political views.

Google’s AI didn’t just lie — it built fake worlds to make its lies look real:

• Fake victims
• Fake therapy records
• Fake court records
• Fake police records
• Fake relationships
• Fake "news" stories

It even fabricated statements denouncing me from President Trump, @elonmusk
and @JDVance
over sexual assaults that Google completely invented.

One of the most dystopian things I’ve ever seen is how dedicated their AI was to doubling down on the lies. Google’s AI routinely cited fake sources by creating fake links to REAL media outlets and shows, complete with fake headlines so readers would trust the information. It would continue to do this even if you called the AI out for lying or sending fake links. In short, it was creating fake legacy media reports as a way to launder trust with users so they would believe elaborate lies that it told.

blablarg18

Judges used AI & didn't check its hallucinated cites.

https://www.judiciary.senate.gov/press/rep/releases/grassley-releases-judges-responses-owning-up-to-ai-use-calls-for-continued-oversight-and-regulation

Grassley Releases Judges’ Responses Owning Up to AI Use, Calls for Continued Oversight and Regulation

The[ir] orders misquoted state law, referenced individuals who didn’t appear in the case and attributed fake quotes to defendants

https://www.foxnews.com/politics/federal-judges-acknowledge-court-ruling-errors-tied-staffers-ai-use-after-grassley-inquiry

Judges blame their own staff.

U.S. District Judge Julien Xavier Neals in New Jersey and U.S. District Judge Henry Wingate in Mississippi

...said the staff members used ChatGPT and Perplexity without authorization

But that only proves, judges didn't check results. Not even from intern.

Blame the intern!

...a law school intern used OpenAI's ChatGPT...

"My chamber's policy prohibits the use of GenAI in the legal research for, or drafting of, opinions or orders," Neals wrote. "In the past, my policy was communicated verbally

HAHAHAHAHAHAHahahahahahahaha

More likely: He never gaf, never had policy. Until he was caught.