The microflow runs in a system context, to which computer should the file be downloaded? That is a rhetorical question: this won't work.
What exactly is your use case? Why do you want to download a file triggered by a scheduled event?
Scheduled microflows are run in the background, so it wouldn’t know where to download the file.
If you are looking to offload document processing to the background, I think you need a slightly different approach.
Instead of having the Download action in the generation microflow, you just need to make sure the document committed to the database, maybe with a “ready” attribute.
You will then need a way to poll for files that are ready to download. This could be a visibility condition on a download button that checks for this “ready” attribute. When it becomes visible, the action on the button could be to download the file.
Hope this helps.