Even though the Speak activity has a very simple user interface, there are IMHO issues.
First, the keyboard "focus". Usually the focus is in the text input field. When you press Enter, the text is spoken. There is a button next to the text input field. When you click this button, a list of previously entered text is displayed. Now, if the button has the focus and you press Enter, the list pops up. Press Enter again, the list goes away, but the button keeps the focus. While I understand what is going on, I would rather like not need to explain this to my four-year old. Or to a six or seven year old child. I know it's a fact of life on all major platforms, but still. I think it would be great for the GUI to have a keyboard focus on text input fields only. Maybe as the default mode, with the option to drive buttons (and other widgets) with the keyboard when the kid is more experienced.
Second, the German voice cannot pronounce Umlauts correctly (although it can handle e.g. UE instead of Ü).
Third, the popup lists. I'll write about those next time.
I will need to find out where to report bugs (or rather, rants in this case), and how I can help to fix them...