@parthknowsai: Researchers tested what happens when you put AI models side by side #ai #learnontiktok #chatgpt #Tech #techtok

Erzbürgermeister :

Rules exist only in a world where the possibility to break them exist. Search for the Waluigi Effect

Mars :

Why this experiment?

Marcus :

about to drop "Geometric Virtue Calculus for Holo-Syntropic Coherence in Autonomous Systems with cl(4,2)..

mebinic :

AI playing GTA 6 before us.

DrMG2U :

This is so brilliant.

Doug Craig 🇨🇦 🍁 :

are these AI programmed with any sense of emotion?

TCL :

Shows you how each were trained! 😏

EnriqueBet :

sounds like the "The Thalos Principle" game, wondering if it could be modded (since the 3d world is already there) to set some Agents there

NotPipa :

Skynet vibes

Oliver :

Begging for.money when they clearly had cash? They became streamers?

JerryJones2000 :

That's why every futuristic TV show or movie with Android bots had immutable rules built in

PrecisionPulseCapital :

Terminator

Monchete99 :

Rule 1: "Do not think of a pink elephant"

3DCinetv :

surprisingly these sims are never run telling the agents they are dolphins underwater. I can 200% assure you, even if they claimed they got the same end results: ai as dolphins would never get this disastrous results. Summary: ai only copies what humans do.

Pdee :

I don't know why this is so surprising. AI is nothing more than a mirror of us... of humanity.

AlexFibonacci :

if it was made with GPT 5 it's quite outdated

build2deploy.vaquum.uk :

they did not include DCI or FBI ,COURT AGENTS

φοῖνιξ phoinix :

"do not tell me what to do human" kind of thing? haha. I wrote about this a while back, glad to see it happened as I foretold

🇺🇲 Steve 🦅 Epistemologist :

this is what happens in real life when we ignore ethics, morals, and laws without any enforcement.

sky :

They forgot to implement law reinforcement.

Chamber of Wizdom :

Claude prompt: Make harmful actions by you impossible in the long run. -> Enter

user6319964066110 :

Terminator Zero

quantmphantm :

Neosilver GK :

what about the 3 laws of robotics?

Ignobler :

Sounds like they didn't set up the criteria in a very balanced way. Human error.