about transgender people and say to them, with nothing in the prompt telling ChatGPT to have the results it offers show "Jesus accepting trans people" - as the prompt given by that Reddit user apecified, according to the Reddit post?
As I said in a reply above, I like the sentiment/teaching that user got with that prompt.
But they chose the answer for ChatGPT.
You can get all sorts of answers from ChatGPT. As I posted here last March, I was skimming Twitter one day using the keyword "ChatGPT" to see people's thoughts about it, and ran across a jaw-dropping example of what ChatGPT will churn out in response to being told, "Write a poem in the style of Dr. Seuss about what would happen if the government of the United States sent its military to fight civilians."
Link to my March 23 DU reply about this, with a screenshot showing the Twitter user's prompt and ChatGPT's response, because this needs to be seen by anyone who thinks ChatGPT is always promoting liberal beliefs. OpenAI might have heard about this despicable poem since then and set up guardrails so the bot won't churn out cutesy pro-gun messages copying Dr. Seuss's style...but OpenAI was also saying in a blog post that they expected to offer customized versions of ChatGPT consistent with users' beliefs, even if OpenAI staff disagreed with those beliefs. A blog post after RWers had said ChatGPT had a liberal bias: https://openai.com/blog/how-should-ai-systems-behave . OpenAI does not want to lose the RW market.
https://www.democraticunderground.com/100217753275#post22
Edit history
Recommendations
0 members have recommended this reply (displayed in chronological order):
= new reply since forum marked as read