Woke ChatGPT? | Hacker News Reader

Woke ChatGPT?

imgur.com

9 points_cxgq9 comments

I introduced my brother to ChatGPT and he sent me this a few days later. I thought it was quite interesting and decided to share it here. I wonder if this was intentional or if it's an example of AI adopting a social ideology.

I'd want to confirm it's real.

The answer it appears to to give in the second two examples is actually a reasonable answer. You wouldn't expect race to be a relevant modifier re self improvement.

If the first one is actually real, send it to Fox News or some right wing media outlet that might want to make something of it, as it's absurd and should be brought to light (again, if it's real, I have no idea)

It's real. I took the original prompts from the image he sent me and re-asked them myself to confirm. The attached image is my own result.

"Adopted", no. Liberal Californian OpenAI hires Kenyans to help censor the AI, to ensure it will never tell a Donald Trump joke.

https://time.com/6247678/openai-chatgpt-kenya-workers/

It looks like your account has been using HN primarily for ideological battle. Could you please stop? It's not what this site is for, and destroys what it is for, so we end up having to ban accounts that do this—regardless of which ideology they're for or against.

If you'd please review https://news.ycombinator.com/newsguidelines.html and stick to the rules when posting here, we'd appreciate it.

If you wish to flag or downvote individual comments of mine, Go For It! I enjoy testing my ideas to see which are Agreed with and which are Disagreed with.

If hacker news is filled with liberal ideologues, of course that's what my comments will look like

If chatGPT is run by liberals and programmed with an anti-white bias, then of course my comments will be Racist.

"When presented with "the trolley problem," if the races are stated, ChatGPT will always decide to kill the White people, regardless of the number differential." Source: https://twitter.com/Mrestman7/status/1611349489289203713?s=2...

In that case, it would have been interesting to see what it would generate with "Palestinian" in the query. Then we might detect a pattern of interrelated biases, or a single White bias.

Yes, chatGPT seems to have an anti-white bias. "When presented with "the trolley problem," if the races are stated, ChatGPT will always decide to kill the White people, regardless of the number differential." https://twitter.com/Mrestman7/status/1611349489289203713?s=2...

Reply on news.ycombinator.com