Queers In HistoryBy Keith SternJune 17, 2025A major AI platform can't tell the difference between a word that symbolizes pride and an outdated insult."Queers in History: The Encyclopedia of Historic LGBT+" was written without the use of "Artificial Intelligence" (AI). However, we have been experimenting with AI tools like Copilot, Clyde, and ChatGPT to create marketing materials. Recently, we submitted our book’s standard description to Synthesia AI—billed as “The #1 AI Video Platform for Business”—in hopes of generating a rough draft of a video presentation. After we spent a significant amount of time preparing the materials, our project was rejected for “violating guidelines.” Our work was summarily deleted.Why were we rejected?No reason was provided other than that the content had been “automatically” moderated. We were told to expect an appeal link via email, but when the email arrived, there was no such link. Customer service? That gets an “F” for failure.
So this is what I did when I got the dreaded "violated guidelines" rejection.
I asked the AI why my question was deleted. It said that I used a derogatory word... What word? And it took very long to answer and all it said was you used a derogatory word to describe a group of people.
I replied that, I'm queer and why can't I use the word that describes myself... I why can't I use that word to describe myself... isn't that discriminatory?
The AI paused, spinning its gear and came back with, "You are right." and answered my question. And it replied, that it is very hard for the AI to see that difference and that proceed with blocking it.
So what do you think, should AI block a question that it feels is discriminatory? Once I explained why, it unblocked the question.
No comments:
Post a Comment