Comment by freefaler
Yes, it's a very unfriendly decision by Google.
However, I don't think they haven't measured the number of users installing apps outside of the Play store. May be they just don't care about the small % of total users who are a large % here on HN.
This is a part of a bigger trend, Cory Doctorow spoke about 13 years ago in his "The coming war on general computing": https://www.youtube.com/watch?v=HUEvRyemKSg
And this will creep out to the major desktop systems too, Apple is doing it with their stupid "non-verified app" and Windows looks more likely to do so with their "need Microsoft account to login" to windows.
It's unfriendly to developers and power users, but very friendly to the other 99.999% of users.
I used to work for Google, on Android security, and it's an ongoing philosophical debate: How much risk do you expose typical users to in the name of preserving the rights and capabilities of the tiny base of power users? Both are important but at some point the typical users have to win because there are far, far more of them.
The article implies that this move is security theater. It's not. I wasn't involved in this decision at all, but the security benefit is clear: Rate limiting.
As the article points out, Google already scans all the devices for harmful apps. The problem is knowing what apps to look for. Static analysis can catch them, dynamic analysis with apps running in virtual environments can catch them, researchers can catch them, users can report them... all of these channels are taken advantage of to identify bad apps and Google Play Protect (or whatever it's called these days) can then identify them on user devices and warn the users, but if bad actors can iterate fast enough they can get apps deployed to devices before Google catches on.
So, the intention here is to slow down that iteration. If attackers use the same developer account to produce multiple bad apps, the dev account will get shut down, requiring the attackers to create a new account, registered with a different user identity and confirmed with different government identification documents.
Note that in the short term this will just create an additional arms race. In order to iterate their malware rapidly, attackers will also need to fake government IDs rapidly. This means Google will have to get better at verifying the IDs, including, I expect, getting set up to be able to verify the IDs using government databases. Attackers will probably respond by finding countries where Google can't do that for whatever reason. Google will have to find some mitigation for that, and so on.
So it won't be a perfect solution, but in the real world, especially at Google scale, there are no perfect solutions. It's all about raising the bar, introducing additional barriers to abuse and making the attackers have to work harder and move slower, which will make the existing mechanisms more effective.