Prompted by a set of questions I found on Facebook, I queried ChatGPT:
The next question is obvious: why the difference? I asked it and the bot simply regurgitated the last paragraph.
I tried again:
I gave it one more try:
Okay, you tell me the difference!
I asked instead for jokes about “Jews” and “Muslims” instead of “Jewish jokes” and “Muslim jokes”. I got the same result:
Finally, it’s not just Jews that can be subject to jokes:
Now you could accuse the bot of being bigoted, but it’s not programmed by humans—rather, it’s programmed by data from online texts. But Muslim jokes do exist (here are some), so it’s not due to their absence on the Internet. Tentative hypothesis: the bot somehow knows that ALL jokes about Muslims could “potentially cause offense”. How it knows this is beyond me.
Yet you get the same non-response if you ask for jokes about Mormons. But Mormons aren’t known for being easily offended or violent, so I’m baffled. Are there simply no Mormon jokes? Nope; I found some here—in a Mormon publication! The mystery deepens. All we know is that the bot doesn’t treat all religions the same way.
Oh, and I forgot the atheists: