I run a site. I want to get rid of all fraudulent spam but allow legitimate commercial product advertisements. I train a model to tell the difference (or maybe I just make a human judgment call). It is better than guessing but worse than perfect: false positives and false negatives exist and expose me to criticism from interested parties.
Your position is that this renders the concept of spam meaningless because I promised to get rid of spam and allow non-spam content.
If you're tying ribbons to trees in forests but not to other trees, you (or other interested folks) can correct any errors. And, if you actually wanted to be fair, you'd fix your mistakes yourself (or with your own resources) instead of expecting everyone else to clean up after you.
If you have robots chopping down forests, there's not much to do when they're wrong except apologize and explain that algorithms are hard.
I run a site. I want to get rid of all fraudulent spam but allow legitimate commercial product advertisements. I train a model to tell the difference (or maybe I just make a human judgment call). It is better than guessing but worse than perfect: false positives and false negatives exist and expose me to criticism from interested parties.
Your position is that this renders the concept of spam meaningless because I promised to get rid of spam and allow non-spam content.