I know this is an old thread but I'm sharing what I hope is a useful update:
For non XML users, to include autoplay in an XLSform, you just add a column to your survey sheet called "autoplay" and on the row of the question you want to autoplay, type audio, or video if it's video, in that column.
However, this only works for a few questions, and causes the survey to eventually crash. It might work if you only have one or two questions with autoplay. Also, I've never gotten it to work for answer choices (maybe because I was doing autoplay on both questions and answer lists on the same screen). There are a couple threads about it on GitHub https://github.com/opendatakit/collect/issues/2861, but it seems to not have gone anywhere - or maybe they are still working on it. If anyone else needs this function, let the developers know! It's great for ACASI type surveys and low-literacy populations, and I'm sure a lot of projects would benefit from this.