IANAL, but wouldn't Google be acting as the recipient's chosen agent to help him communicate via email?

E.g., what if the recipient has a paid human assistant who screens his email, wherever hosted. Further assume the sender knows nothing about said assistant, but would be mortified if she did?

Google is the recipient's assistant and is compensated by being allowed to show the recipient ads based on an automated analysis of his email's content. How is an automated assistant materially different from a human one?

I don't see the problem.