That’s arguable. We, as programmers or non-, can certainly envision bots that coldly calculate equations dealing with death destruction, images of swastikas, explosions, executions, &c. If we have a news-manipulating bot, should we keep the bot from manipulating these images? Should we bar it from using certain words that may show up organically, but in awkward contexts (n1**er1
or gay or whatever)?
I’ve used the wordfilter module on a few bots, I’ve also whitelisted some of the words.
See also: transphobic-joke-detection
Obviously, some people think about these things more than I do. I am ambivalent.