I even had to change name and function for one of our components in 吃！use of a component called gun, was not acepted at all, particularly when the ptotagonist character, aims it at his mouth twice!! even if it is to shoot candies!!
So we substituted the gun name for a 'bang toy", engraved with a number 2, with a mouth shouting bang every time it is filled with candies! You can only refill this bang toy twice, after the second time the number two dissappears and the mouth cannot have any more candies. So kids who own the device cannot eat anymore.
This version of the mnemonics narrative was approved. The resulting image is an excellent memory hook, at least for me.
Lol - that’s quite annoying. I get it, but now you point it out it’s a bummer. If I ever started using AI generated images none of my mnemonics would get through because they’re all either R or XXX rated. It’s annoying how well adding sex and/or violence to mnemonics works … obviously HH can’t do that for the public ones by default, but for anyone ever struggling with remember a word I highly recommend it if you’re not trying to get it past ChatGPT filters XD
Now I cannot create anything mentioning Mickey mouse, even by using the black and white edition, they will not badge. I ask ai to create a cute doggy with big mickey mouse ears, but that did not work either. I request it to keep on trying and adjusting until it can produce something.
I have to rethink ‘mi’ from Mickey to what?? Minty sounds good, is that copyrighted?
Sorry Mickey Mouse, you have been pushed out of my mnemonics cast. No worries, your spot has been properly taken care of: a perky, fun loving and cute goblin, elfy type little princess emerged, summonded by the depths of AI dimension. Her name is Mitty and has taken ownership of ‘mi’ initial, , with her first hanzi, appropriately meaning ‘name’: 名
It seems like using historic, iconic building names also raises red flags with AI moderators. We had to replace ‘English Manor’ with a descriptive prompt like “luxurious rich mansion” to get it through.
The anoying thing is that we only find this out by trial and error. They do not come back explaining why the prompt was rejected, they just say ‘it has triggered content violation flags’.