For anyone who's used Google Translate, it should be entirely unsurprising that the best "AI" we can cook up appears to be doing little more than rating individual words.
I admire the effort, but the tools seem to be so fundamentally lacking.
Give it some time. ML improves with usage and training data.
While this is true, ML isn't a panacea that will solve all problems given time and training data.
The same quote is only 12% "toxic" if you remove the ... inside, and only 10% if you quote the entire thing (with ").