My guess is it would probably take resources SqE don't have or as a company gain very little (to them) to use. It's a machine learning problem where someone / a group will have to first sit and manually flag several millions/billions of examples of chat messages as RMT or not, to train and run a Bayesian network or support vector machine algorithms (or RNN, FCN, etc... i.e. different AI neural network archetypes) on.

(I realize by writing this that it can be misused) If the bots break up the lines it becomes much more difficult for an algorithm to recognize and block the text real-time, same with adding spacing b e t w e e n characters, symbols, and adding non-printable (invisible) characters. Human read-able, but more difficult for algorithms that build up/use dictionaries from the training. Then again, one could design the algorithm to look at the full chat history (with/-out spacing) after being reported for RMT. Then the botters would have to figure out how much they could spam before being banned, and such, and in some cases have to intersperse with other text than spam. At Google they regularly change the algorithm up to prevent misuse of the search ranking. The developers could do the same e.g. each maintenance day, but compared to Google, they don't risk much by having the bots, and the botters sadly lose little by being blocked. Maybe they already do something like this when doing their ban-waves, but who knows...