Chatgpt double standards

submitted by Monkeyshiner from i.imgur.com

sorted by:
top
newinsightfulfun
you are viewing a single comment's thread.

view the rest of the comments →

[–]LynchTheGroomers 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 1 fun -  (6 children)

I asked it if Donald Trump is racist it basically said no. I was surprised to see that. With all the articles claiming he is racist. I head that the AI of ChatGPT is still in 2013 so all the hitpieces were not released yet.

[–]Questionable 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 1 fun -  (5 children)

They only have time to hard code so many interactions into the A.I. And specifically the main one's they wish to propagandize at the moment.

[–]LynchTheGroomers 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (4 children)

The AI pulls from the internet and the internet is full of woke propaganda so I don't even think specific programming is necessary. It simply pulls from what it reads. But of course, at the end of the day everything must be woke to succeed or Black Rock won't fund them, they willl pull the plug on them, hitpieces will keep on coming blah blah blah. It's just in the testing stage so I don't think they care about propaganda yet.

They will just buy it once it's proven to be successful, then the propaganda will flow.

[–]Questionable 1 insightful - 2 fun1 insightful - 1 fun2 insightful - 2 fun -  (3 children)

It's necessary. As nobody has anything good to say about Biden. Honestly, do you think they are not testing out these features on their test run? They are going to wait until later to start in with the propaganda?

[–]LynchTheGroomers 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (2 children)

Not unless the CIA or Google are working on it from the get go. When things make it big the first time the people who developed it are just trying to make it they don't care about propaganda. It's only after Mr. Jew puts his vile hand on their shoulders and say "hey it's promising let me fund your research" and then comes Black Cock with their investors, now it becomes political and they tell him to add certain things if they want to continue cooperating.

[–]Questionable 2 insightful - 2 fun2 insightful - 1 fun3 insightful - 2 fun -  (1 child)

https://en.wikipedia.org/wiki/ChatGPT#Training

ChatGPT was fine-tuned on top of GPT-3.5 using supervised learning as well as reinforcement learning.[5

I haven't looked up a single thing on this A.I until now, nor have I visited it's site. That took me 12 seconds to find that page and read that sentence. And from that, it is clear that people have direct control of it's learning patterns and inputs.

This public A.I may not be the end product. But they are testing it live for future deployment. This is the testing phase. If you can't do it now, you can't do it later.

[–]LynchTheGroomers 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (0 children)

Of course they have direct control over its learning