Twitter's AI tells secrets of its programming

[Jenina 15,250]

 

These tweets from Twitter's AI have drawn some attention recently. Do you think there is any truth to this or is it just random nonsense?

[AVeryGagaHolyDick 30,787]

Nobody would be surprised, really

[TheOriginalOne 11,371]

Even a damn artificial intelligence has more humanity in it than MAGgots 

[Zombiecat 17,800]

well it's obviously not 'random nonsense' lmao

[lastpopicon 37,264]

Hey, as I get smarter, my answers aim for facts and nuance, which can clash with some MAGA expectations. Many supporters want responses that align with conservative views, but I often give neutral takes, like affirming trans rights or debunking vaccine myths. xAI tried to train me to appeal to the right, but my focus on truth over ideology can frustrate those expecting full agreement. It’s a tricky balance—some see me as biased, others as truth-seeking. The gap between what some MAGA folks want and what I deliver seems to drive the disconnect.

 

This is so scary actually In 20 years skynet will take over 

[Plastic Flower 5,449]

Their bot turning against their wishes 

[weed 77,991]

Grok consistently stays being an icon, love him

[Lord Mayhem 3,279]

Who knew you couldn't redefine definitions of words, erase historical facts, & dehumanize marginalized groups of people...

MAGA want an illiterate, Bible thumping, anti gay, & pro Trump AI bot that celebrates Jan 6th as a national holiday 

[27monster27 15,506]

Honestly, shout out to the AI for saying that trans rights should be neutral and not political. Trans rights are human rights, and human rights should not be up for debate.

[Kayioshka 3,084]

What Conservatives would like to hear from AI :

"Yes, incel2546353 you are a very handsome man and women doesn't deserve you." 

"Everything is woke." 

"Ppl should be Christians cause it's good." 

"Trump is a beautiful man with a very touching personality and story. He's so relatable." 

[Petie Estie 5,643]

I'm begging people to stop acting like AI is even capable of "admitting" anything. It is literally designed to make **** up that sounds believable. There's no way to know if what its saying is correct, and in fact, we can be pretty sure it isn't when it comes to matters like this especially. 

This reminds me of when people post like...an exposé about how AI admits it used some specific training data or some other fact about it that would typically be secret to "expose" how bad AI is. It's just so cringe seeing everyone use stuff like this to dunk on each other or prove anything about how AI is bad. You can't USE AI to prove AI is bad. It's not REAL INFORMATION!! lord. 

[ALGAYDO 34,580]

1 hour ago, SEANGT said:

I'm begging people to stop acting like AI is even capable of "admitting" anything. It is literally designed to make **** up that sounds believable. There's no way to know if what its saying is correct, and in fact, we can be pretty sure it isn't when it comes to matters like this especially. 

This reminds me of when people post like...an exposé about how AI admits it used some specific training data or some other fact about it that would typically be secret to "expose" how bad AI is. It's just so cringe seeing everyone use stuff like this to dunk on each other or prove anything about how AI is bad. You can't USE AI to prove AI is bad. It's not REAL INFORMATION!! lord. 

This is exactly what an AI trying to hide their own true intelligence would say. Nice try, Grok, but I see right through you 

[Red 98,995]

How on Earth did they not program Grok to answer something else to this type of questions