What you really need isn't a breakup, it's some way to foster sustainable competition for search.
Google is programming people on a massive scale. Compare your news feed to your partner/friend/colleague’s phone.
No-one I know outside of tech has any notion that what they’re seeing on that feed is only shown to them. Social targeting in tech has completely broken society with these tools - and Google is arguably worse than Facebook in this regard because their feed is ubiquitous, people just don’t realize they’re in Google’s world the entire time they’re on the phone because they didn’t touch a Facebook or Instagram icon to start an app.
We assume non-tech people out in the world don’t know much about how this works, but the average user is even more naive than any of us realised.
Would it be “better” (given the regulatory goals) to order Google to do XYZ and provide an audit trail to demonstrate compliance?
Putting pressure on Google requires seriously increasing competition.
Let the local units rebrand and compete against each other. Maybe even auction them off to other big players. Whatever needs to be done to even the playing field. Google need a nice kick in the ass for their own good health.
What actually created competition for the telephone network was the internet. Competition from cable and satellite and cellular.
That's what you need. Something similar but not exactly the same, so they can each carve out a niche and provide a backstop against the most flagrant abuses but they aren't in such direct competition that they're more inclined to cooperate than compete. So that you have companies who can make choices knowing that their position is different than their competitors and the competitor can't just mirror their action to neutralize its competitive advantage, which otherwise removes any incentive to be the first to engage in pro-consumer behavior.
Being able to simply dictate who can do what on the internet solely because they can undercut everybody's price (aka give away email free in return for mining your data) is far too much power.
Are those mails marked as spam? Or what happens?
As I am the sole person to ever use this server and I have been sending to some of those emails since almost as long as gmail has even existed, the fact that they are now marked spam is simply an admission that Google doesn't actually give one shit about accuracy and solely trains their ML networks on things that bring them revenue.
This isn't specific to Google: all the companies that receive email have the same difficult problem. Email is the only major fully open federated push protocol, and spam has come close to killing it.
(Disclosure: I work for Google, nothing related to mail, speaking only for myself.)
If the trade-off is that Google accommodates mails sent from badly configured servers and I receive a lot of SPAM, or they drop that mails and limit SPAM, I will go with less SPAM. I know it's not cool for decentralized Internet, but that's the utilitarian reality for me.