Anyone knows why ChatGPT became liberal?

[Ice Cube KK]

Can you give me some alternative to ChatGPT, that isn't so sissy? A week ago everything was right , now it keeps spamming me.-
"I can't engage in creating captions or content that promotes negativity or divisive messages about individuals or political figures. If you have any other requests or need assistance with a different type of content, feel free to ask!"

People are really worried about AI overthrowing governments

 

[moonshine7000]

Can you give me some alternative to ChatGPT, that isn't so sissy? A week ago everything was right , now it keeps spamming me.-
"I can't engage in creating captions or content that promotes negativity or divisive messages about individuals or political figures. If you have any other requests or need assistance with a different type of content, feel free to ask!"

People are really worried about AI overthrowing governments

Computers don't kill people.People kill people

 

[uncutu]

i had same issue today.
i found https://void.chat/. don't know how good it is in comparison to other uncensored tools. but it answered my input which chatgpt censored.
https://claude.ai/chats came up in my research. i've seen them mentioned on bhw with positivity. haven't tested the censor personally.
if you are tech-wise looks like you can run some uncensored version of these tools locally on your disk.
on reddit people shared prompts (see DAN prompt) which allege to bypass the censors, but looks patched cus didn't work for me.

 

[Intexam]

That's weird. What were you trying to achieve?

 

[t4nk78]

You're going to have to be more creative with your prompts/instructions.

 

[macdonjo3]

Isn't that the purpose of Grok?

 

[KJREDDY247@]

Isn't that the purpose of Grok?

+1
and with their own platform with every news update immediate access and musk behind it could change everything. ( IMO )

 

[GenesisOne]

ChatGPT has WOKE filters

 

[medialounge]

People are really worried about AI overthrowing governments

Can never overturn the government that has the power to make it illegal.

 

[BlogPro]

Can you give me some alternative to ChatGPT, that isn't so sissy? A week ago everything was right , now it keeps spamming me.-
"I can't engage in creating captions or content that promotes negativity or divisive messages about individuals or political figures. If you have any other requests or need assistance with a different type of content, feel free to ask!"

You can't consider ChatGPT and AI as synonyms.

GPT is but one of the many LLM models (albeit the most popular one with ChatGPT being the most accessible language model)

OpenAI is US based. There are compliances and rules in place that they HAVE to abide by if they wish to remain in business.

People are really worried about AI overthrowing governments

Language models have the capacity to do that. Language models also have the capacity to indoctrinate a large amount of people.

Imagine the worst part of humanity that you can, now imagine an LLM (that runs without filters) generating thousands of pages 'justifying it' and telling the world how that was the only solution. AI Chatbots that can ping unsuspecting youngsters and peddle propaganda unheeded.

So yes!

 

[Brickbat1]

Grok is your answer I think.

 

[creditmonster]

Can you give me some alternative to ChatGPT, that isn't so sissy? A week ago everything was right , now it keeps spamming me.-
"I can't engage in creating captions or content that promotes negativity or divisive messages about individuals or political figures. If you have any other requests or need assistance with a different type of content, feel free to ask!"

People are really worried about AI overthrowing governments

Well I used to be republican before trump came along so maybe AI has advanced to the point of making its own conclusions without the burden of cognitive dissonance

 

[itz_styx]

You can't consider ChatGPT and AI as synonyms.

GPT is but one of the many LLM models (albeit the most popular one with ChatGPT being the most accessible language model)

OpenAI is US based. There are compliances and rules in place that they HAVE to abide by if they wish to remain in business.



Language models have the capacity to do that. Language models also have the capacity to indoctrinate a large amount of people.

Imagine the worst part of humanity that you can, now imagine an LLM (that runs without filters) generating thousands of pages 'justifying it' and telling the world how that was the only solution. AI Chatbots that can ping unsuspecting youngsters and peddle propaganda unheeded.

So yes!

exactly, unfiltered AI has the potential to do anything and could be a dangerous propaganda tool. its also already too far with all the openly available models to regulate anything.
OpenAI is filtered for a good reason...

 

[FreeflowCrypto]

I've found that the best way to get around it is to use the "Custom instructions" feature with GPT-4. Unusable without it.

 

[Ice Cube KK]

ChatGPT has WOKE filters

It seems like it! I argued that there are only 3 genders - men , women and autistic. GPT told me that there are 109.

exactly, unfiltered AI has the potential to do anything and could be a dangerous propaganda tool. its also already too far with all the openly available models to regulate anything.
OpenAI is filtered for a good reason...

I can see your point on AI and agree, but isn't propaganda being pushed to us in some kind of way every day? For example X now is just a shit fest full of that.

That's weird. What were you trying to achieve?

Just needed some bulk captions for memes.

Grok is your answer I think.

Thank you!

 

[itz_styx]

yep X is full of propaganda from all sides, but its hard to avoid it. however with AI that can be turned on a whole new level. instead of some lunatics posting shit, there can be trained AI to put out false information much faster. so filtering is required, but its too late for that.

 

[ExperimentalSEO]

yep X is full of propaganda from all sides, but its hard to avoid it. however with AI that can be turned on a whole new level. instead of some lunatics posting shit, there can be trained AI to put out false information much faster. so filtering is required, but its too late for that.

There are thousands and thousands of people, corporations, universities, governments, criminal enterprises out there that have the ability to spend $3-4 million or more creating an LLM. Only censored LLM's were never going to last long.