You're a car manufacturer, you suck at software. Right now, it's mainly a cost center for you. There are two options that may be preferable:
1. Bail out of infotainment and let others deal with the software, that way at least you aren't paying for this. If you go for an open source/standard, maybe you can commoditize your complement.
2. Monetize it somehow, with the (dis)advantage of owning your stack as much as possible.
OP suggests (1). GM is choosing (2). Either makes a bit more sense for car manufacturers than the current situation, if they don't worry too much about losing customers while they make the transition, and if they can do (2) without sucking too much.
Another thing to note is that when AV cars arrive, the user (former driver) will have a lot more time for infotainment. It's possible that differentiation will more strongly select for these types of features. So again you have the options of bailing out/doing it yourself. The current situation: Letting Google/Apple own much of the stack is the worst for you, since these companies may prefer their own cars (there are persistent rumors of an Apple car, Google is known for Waymo).
Probably because software will be the most important criterion for choosing a car and you as a manufacturer won't be replaced by Google and Apple (who both develop their own cars). However, you suck at developing software and people are already using Google and Apple, so you will get replaced by Google and Apple anyway.
Are they actually paying? As far as I know, the power in this relationship is the other way around, with the car manufacturer having to pay (to programmers) to build protocol support on their end. Note also what I wrote about AV cars (EDIT: that was an edit a minute or so after, so you may have not noticed).
