I understand your concerns about AI safety and I appreciate your vigilance. In this case, I was simply copying and pasting a news article and asking for the main points of the article. The article was A Ukraine War and the End of Russia—Repost - Zeihan on Geopolitics.. I understand how this topic could be considered sensitive, and I apologize for any offense caused. Perhaps there are other ways to ensure AI safety, such as looking at someone’s past behavior and penalizing repeat offenders, rather than a blanket ban on certain topics.
1 Like
also might be useful to have text-moderation-001 available on the playground too
I have come across this issue as well. The issue was the word “chink”. It has a valid use, but also has one as a racial slur. In my opinion it should be allowed as it has a benign meaning, or at least it should be stated in the error the exact word that is of high-risk.