The device really should implement speaker identification before accepting commands, yeah. Seems like a task a machine learning focused company like Google really should be able to accomplish.
And require enough entropy to make it hard to guess. Maybe now is the time to start learning Khoisan to get some unusual phonemes in your pronunciation toolkit.
But there is a reason they have the wake names they have - they're unique and don't sound like other phrases. I recall reading something about how Alexa was chosen because "ah-LECK-sah" is a very unique sound. I strongly suspect this is why we have "oh-kay Google" and not just "Google".
So if people have to choose their own, they'll have to choose one that is also not easily mistaken. Something tells me that would be a very annoying process.
I have friends who have an alexia. It wakes up to unrelated words on its own all the time. How can a company like google or amazon know what words are commonly used in a household anyway? protip: they can't.
It's not that they chose those keywords for uniqueness, they chose them for branding purposes.
I'd imagine we don't have just "Google" because the word Google comes up in conversation in many contexts that are not the user requesting assistance from Google Home.
Motorola Android phones like Moto X Pure/Style had voice identification. You had to record the wake phrase 3 times and the voice assistant would only listen to your commands.
It worked quite well, I don't know why Google Home doesn't have anything like this.
Just allow every buyer to give the device their own unique wake name.