* Website gives person a unique code to broadcast on the blockchain to prove they have control over that identity
* 'Approved' services claim things publically about identities (eg age, allowed to drive) which others can use to verify claims
What I don't get is:
* how you know which 'approved' services to trust in the first place
* why you'd want all of that information public
* how it'd fix the original issue
The other issue is that none of this is necessary. If I want to prove to Person X that Gov Y says that I have property Z, why can't I just present a cryptographically signed certificate saying I have property X. You could even have a similar revocation list as CA.
You don't think if the problem were that easy to solve, it would be solved already? Throwing blockchain in the mix gets Facebook nothing that it couldn't get today by just partnering with organizations can verify identities.
Not to mention, the "fake news" issue isn't largely an identity issue — there are plenty of people and organizations willing to lie on their real accounts to further their sociopolitical agenda, and plenty more people who don't care if something's true so long as it confirms their existing viewpoints.