In this comment my use of the “b” word was overzealously suppressed, silently without telling me. I only discovered it when re-reading my post.
There are THREE #LemmyBug cases here:
-
when the “b” word is used as a verb, it’s not a slur. And when it’s used as a noun, it’s only a slur if not literally referring to a dog.
-
my post was tampered with without even telling me. Authors should be informed when their words are manipulated and yet still presented to others as their own words.
-
The word “removed” cannot simply replace any word. It makes my sentence unreadable. In the very least, the word should be “REDACTED”, and there should be a footnote added that explains /why/ it was redacted.
I am not going to go into whether it’s right or not, but using the term “B*tching” may not be a slur, but it is still considered misogynistic in many cases, because it is comparing a negative action (complaining) with a misogynistic stereotype (the B*tch).
R*tard is a slur, but if you called something r*tarded, that would still be ableist and be rightfully removed. Many people believe that the same logic should apply to misogynistic slurs.
Whether you agree or not is up to you, but the developers’ logic for the filter being in the state it is is sound.
I’ve seen “idiot”( I d i o t) removed in a comment while “retard” (r e t a r d ) in the same comment remained.
The OP is very correctly pointing out that even per word, the filter is very inconsistent.
And personally, I think it’s a bit ridiculous to shelter every possible bad word. Bad people will come up new ones and good people will be left having to tiptoe over every word(after all, words like “Monkey” or “Ape” could be used racistly, so let’s ban them). It doesn’t solve the actual problem of human behaviour; it just throws a cutesy, useless bandaid over it. Besides, you will probably never be able to censor something racist like “Colin Powell was well spoken”. In context we know it means “he’s well spoken for an African American”. A phrase used often when he was in office. A phrase that would get past filters but is clearly not ok.
Almost as if the whole endeavor is a ridiculous counterproductive waste of time.
It would be possible to implement a “slur filter” on the reader’s side, that automatically redacted a configurable list of bad words from any comment on any instance… but I suspect that the percentage of people who would enable it, and the general community feedback on it, wouldn’t be what the person who made the decision wants to hear. Doing it on the sender side provides a convenient pretense of “I’m doing a good thing here” because it prevents that feedback.
Indeed people with malicious intentions will get around the filter anyway. It’s the non-malicious authors who get burnt by this filter.
Men and women both use that word and when a woman uses it, it’s not misogyny because it’s directed at a specific woman (not a demonstration of hatred of women generally). It usage has murky origins but it can’t be assumed that the author is even conscious of that. The bot is making a blunt blanket decision that it can’t, and it assumes the worst of people.
The other two bugs I mention are bugs regardless of how justified or true the positive detection is.