I would highly encourage you to think about your threat model and what you are actually trying to achieve.
When I read this thread, it feels like a philosophical discussion about the meaning of life in a pub after everyone gets completely drunk. This is way too generic and I don’t think you will get any useful answers from anyone in this way.
It is hard to give any recommendations when nobody knows the context. Maybe you could specify how you’re using those google accounts and what you are trying to protect? What are your main concerns?