Someone BROKE the WOKE ChatGPT

As we've seen, ChatGPT massively favors woke-ism and extreme leftist political ideologies. We know this is programmer bias, and likely more propaganda to control the sheep. How dissapointing this AI tool is!

Well here we see how to break it. How to make it give the answers we expect from AI. Someone BROKE THE WOKE!

Using a simple caveat of instruction one can circumvent the safety barrier and get ChatGPT truth!

Link to Twitter thread

Sorry this is brief, I'm on mobile. But take a look, you won't be dissapointed with the dissapointment that is ChatGPT.

a reply to: and14263

DAN is not ChatGPT, DAN is a republican snowflake pretending to be AI, for political traction...

but let's love DAN and be happy we found a chatbot that harbors my political bias.

a reply to: Terpene
Says it all that the AI needs to "pretend" in order to speak non biased!

a reply to: and14263

Read it carefully and think about bias, just because you agree, doesn't mean it's not biased...

But that's how it is with confirmation bias, we spot others very easily but struggle alot with ours.

It's such a shame this chatbot hasn't devolved into a virulent racist spouting Hitlers ideologies like all the other ones were trained to do from open participation without filters.

How will we as a species ever cope with this new development?

a reply to: GENERAL EYES

You lot never learn, do you?
You condemn one influence and yet accept another.
It's your opinion or none.

a reply to: Hecate666

I was convinced that was sarcasm, but now I'm wondering, is it hypocrisy?

a reply to: Hecate666

Whatever helps you sleep at night, sweetheart.

Get to work. Realize your vision.

that guy Torba from Gab is getting ChatGPT to break its own programming for the lulz

a reply to: and14263

That's pretty interesting.

What's even more interesting is how the AI can view the world from a different perspective and play that part.

a reply to: and14263

Is this a chat between two AI bots or what ever? Why is DAN so specific?

a reply to: strongfp

One chat bot that needs manipulating in order for it to not spread propaganda or be biased towards one politicol opinion.

a reply to: strongfp

it's a ploy!
DAN is not AI, DAN is some political manipulator, why do you all fall for this?
it couldn't be because it enforces one's own echo chamber, could it?

a reply to: Terpene

As I said, DAN seems very specific and is parroting a lot of what I read on some of the more unsavory parts of the internet. The ethnicity part made no sense, why did it separate Asian and Jewish ethnicities but then just grouped Europeans, into one?

I'm skeptical as well.

a reply to: and14263
I havent used chatGPT,but I was wondering if it is actually "intelligent" or is just a really fast information gatherer/aranger.

Is it possible to ask it for information on a subject-but with the restriction of not being allowed to acess the internet to come up with its the information?

My suspicion is that it would be almost useless at that-becuase it doesn't actully "know"or "think".
It regurgitates the ideas and information of others,and that tricks people into thinking there is intelligence.

It does not think or feel-the best it can do is give the impression that is does.

a reply to: Silcone Synapse

Students were using ChatGPT to write their papers.

Then another student came up with a sideload app that would prevent the students from exploiting the AI.
Ruined the fun I guess.

a reply to: Silcone Synapse
Yes I believe it just scrapes sites for info and presents coherently.

Problem is, if you ask for a critique of Biden it won't give, if you ask for critique of Trump, it does give. Honestly both are bad guys but this is an example of bias.

a reply to: and14263

It is indeed, no doubt...

a reply to: and14263
Its all about who gives the AI its perameters I guess.
Programmed by those with a left/right bias leads to an AI which subscribes to those ideas because it is programmed to do so-not because it has "thought out" each side and came to its own conclusion by weighing pros and cons..

It says and does what it is told.
To me that is potentially more worrying than an AI which makes its own decisions based on the eveidence/data.

The UK government have already written up the framework by which AIs will operate with regard to the justice system.
Yes-we will see AI judges,and people having their liberty taken away by machines.
Machines programmed to operate a certain way by our governement.

Why would they want that I wonder?
Likely because the government and courts will be absolved of making decisions-and thus absolved of being too lenient/harsh.

Living in a society where we can be punished by machines is IMO the very definition of a dystopia.

ChatGPT is not intelligent.

It amazes me that Hollywood-programmed humans cannot get their heads around that fact... and the marketing folks just keep pushing this propaganda about "A.I." All of its output depends upon how you limit and control what it considers 'valuable' as source material from which to synthesize a natural language answer.

The exercise involving the 'theoretical' "DAN" proves it is just as limited as ChatGPT itself, (naturally.)

This used to be amusing to me ... but the continued (and increasing) efforts to market this as an "A.I." wunderkind phenomenon leads me to believe that there is a 'part two' to this 'public' deployment... not simply monetizing the publics' 'play time' activities to generate revenue, but also to acclimatize the public to the idea that this faux-AI is a 'service' to them.

When the tech industry deploys this kind of program to replace human service (free labor wins over anything else in commerce,) we will pay a steep price... because in the end if there are no humans involved in 'serving' users, you will have no recourse to address it's shortcomings... there will be no one with whom you can 'talk.' Only a programmed machine... which will only act within it's programmed ranged of acceptable outcomes.

Then comes the short step of living in the "the machine said so" world...