https://www.theverge.com/2018/2/1/16959454/voynich-manuscript-mystery-ai-decoded-debunked
Hmm...
You would think that 25-30 letters which is maybe standard(?) in most languages? Aside from Chinese(Mandarin) and some others. Comparing it to over 300 languages and looking for frequency of characters would yield something. The article makes a good point that they used todays languages against a manuscript dated to 600 years ago. Kind of how people interpret the bible today based on writing from 2000 years ago, but I won't get into that more here.
Using AI is a good step because you can conduct many lifetimes worth of experiments and comparisons. I do find it interesting that they said it was full of anagrams. Also being a voweless(*) language that they claim it to be allows for so much subjectivity in the final analysis. I also IRL LOL'd when it talked about using Google Translate. I mean, if you are submitting a scientific paper and using Google Translate, I mean come on.
Chings have you put this script through CHaMOY or maybe
@mocat could put it through SNOWBRAG? I mean anything could decode this, maybe it's the batheing championship rankings from 1463. Regardless this is very interesting.