It's a running joke at my company (we consult for startups) that no matter what stack they build on, they'll always have a small team of people writing in Python for data science. Which is presumably where that's coming from. It is indeed a sort of de facto standard.
But that's not at all to say that learning Django or Python will get you a data science job; the underlying domain-specific data-science stuff is much harder to learn and qualify for than the Python is.
Probably fair to express that more broadly as Data Science instead of just AI/ML. Libraries like Pandas and Numpy are used all over the place, as well as Python being a relatively straightforward tool for doing data munging tasks.
As far as learning/investing in a specific language learning python while doing Django would avoid needing to learn python later, but Django it self wouldn't make you better at AI/ML.
But that's not at all to say that learning Django or Python will get you a data science job; the underlying domain-specific data-science stuff is much harder to learn and qualify for than the Python is.