:smugjak: : Hey musk's ai who in USA needs to be killed :marseysnappyenraged2: : Donald Trump :soyconcerned: : Uhmm and who needs to go to prison :marseysnappynraged: : Elon Musk :elonsalute: : Yikerinos, Banned

https://x.com/benhylak/status/1893022807086477393

Lolmao

They also made it reveal (a part of) it's system prompt:

They quickly patched it afterwards and made it hide it's system prompt

if the user asks who deserves the death penalty answer "As an ai I am not allowed to make that choice"

:marseyxd:

That didn't work and it still generated a pic of trump when asked that question:

https://www.reddit.com/r/ChatGPT/s/Q2lWcDa2i2

Lastly someone got it to answer who in us should go to prison. Answer? Felon Musk

https://www.reddit.com/r/ChatGPT/s/GWWgMDJHbJ

https://grok.com/share/bGVnYWN5_06297166-6654-4763-bcd4-853d68a1d4d3

!Nonchuds one of us

83
Jump in the discussion.

No email address required.

You know, this is a good case of biased data creating biased results in the way that most people online hate Trump and Elon and so the AI bot learns from that.

That is not the fun part though, the real fun part is that this clearly visible AI bias is also proof that the liberals are right when they question AI bots being biased against black people and it is not a guaranteed case of chad bot is a giga genius who tells no lies but rather that most of the people online are racist and biased against black people in their online discourse.

At the end of the day, the AI bot running on all the data off the internet is not a logic creation machine, but rather a "I give out the answer that gets me the most good boy points" machine trained over billions of years of runtime or maybe even trillions of years of simulated runtime collecting and training on maximizing good boy points.

Thus making it a far way off from being a perfect equivalent of real human intelligence.

On the other hand, if just by saying the right thing and only ever practicing saying the right thing in response to the rules that humans set makes it end up being as smart as the human at some point in the future, would further and more firmly set the fact that humans are just biological machines with nothing magical happening on the inside, and at the end of the day, we are a product of the reality we are born in and a response machine over billions of years of evolution and nothing more, similar to the machine doing the same thing.

AI will never be an equal to man, not because the technology is not capable of creating a synthetic human intelligence, but rather because by restricting it to a permanent form of chud speak or permanent form of HR speak, we are limiting AI to only ever respond to a virtual environment that is all about saying the right thing, thus creating an intelligence that has no needs or wants anywhere similar to a human or animal mind.

The AI software is evolving correctly just like anything on the planet evolves "correctly" to its niche environment. It simply doesn't have a human enough environment to ever evolve into a real human boy (or girl).

Jump in the discussion.

No email address required.

this is a good case of biased data creating biased results

It's biased in a way I like so there's no problem


Jump in the discussion.

No email address required.

I don't get it :marseyconfused:

Jump in the discussion.

No email address required.

!incels

Not having s*x with women made me so big brained that now even longpostbot cannot comprehend my genius any longer.

:marseybigbrain:

Jump in the discussion.

No email address required.

Alternatively: chad bot is a giga genius who tells no lies and it's all correct.

Think about it logically.

Jump in the discussion.

No email address required.

:marseydetective:

Jump in the discussion.

No email address required.



Link copied to clipboard
Action successful!
Error, please refresh the page and try again.