That’s more like statistical inference, applying insights from known languages to it. It’s not really ML. Also, it can give you some broad clues, but decipherment is still a massive leap.
By and large, ML is statistical inference albeit in less-than-rigorous form.