Data portability is easy to implement (and has been by some companies).
But back-end interoperability and delegability are tricky not just because of technical challenges (those can be solved), but mainly because of abuse and spam.
If, all of a sudden, any other service can like and comment on Facebook posts without having a Facebook account... I don't see how this doesn't become a spam/abuse-fest overnight.
My $0.02 is that there's a much simpler solution hiding in plain sight: just stop giving legal copyright/hacking/ToS protection to sites with third-party-generated content that the site didn't license.
In other words, if someone wants to build an app that downloads your Spotify playlist data by impersonating you... or scrape Craigslist listings... or copy Yelp reviews... then nothing is legally in the way. Not copyright, not anti-hacking laws, not ToS.
Services are free to rate-limit or ban you if they catch you... but if you can get around that then the information/interaction is yours.
This feels like a much more "free market" approach. It doesn't rely on trying to build laws that will be bad compromises and outdated by the time they're passed. Rather it involves undoing existing laws and protections that are producing harm.
(To be clear, copyright still applies to things that are paid for or licensed -- movies, songs, articles, etc.)