It depends. If you intend to use this with an actual baby, you may want to think about whether it's reasonable to use VR at all - it just isn't all that reliable. An IR remote might be a better choice.
If you're building something for fun or as a demo for school, it looks like the VR works some of the time and commands get through to the Arduino. So you need to spend some quality time debugging to find out why the motor doesn't spin up. Divide and conquer would be a useful way to go there. But it will never be flawless - you could find that the best thing you can do is to train yourself to speak more clearly when giving commands.
There are Arduino projects on-line using Alexa as the VR component and I expect there's stuff for the Google version too. Those would likely get you better accuracy, but I still expect you'll find yourself repeating what you said, sometimes multiple times. I certainly did before I turned Alexa off.
Whatever route you choose, there's going to be debugging though. If you're expecting it to be plug and play, I suspect you'll be disappointed.