I don't think this model is within the realm ML (it's theoretical neuroscience; although there is much overlap) but does qualify as AI which is what was asked about in the post title.

There is an annual symposium called the International Workshop on Active Inference for about 3 years now where research is presented and the papers are linked there.

And of course the dozens of research papers you can find through Google Scholar on the topic.

Edit: I did find where a few active inference papers have been presented at NeurIPS.