"AI companies are playing up extinction risk for marketing" reminds me of a different take I've heard multiple times:
Coca-Cola introduced New Coke, which was widely disliked, and then they switched back to the Classic Coke formula. I've heard multiple people say that New Coke was supposed to be bad, and the secret true reason was so they could switch the normal Coke formula from sucrose to high-fructose corn syrup without anyone complaining.
My response to both of these is the same: no, obviously this was not a marketing scheme. Companies do not risk the future of their business on purpose as a marketing trick. Coca-Cola switched to New Coke because they genuinely expected people to like it better. AI company CEOs talk about AI risk because they are genuinely concerned about AI risk.
I think it's also marketing but not necessarily to the rest of the world, but to their employees.
It's saying something like "what we are building is so powerful, it can even wipe out humanity. You and me, we are the most important people in the world right now commandeering this immensely powerful technology. Haha us nerds finally on top of the world. And look! We are the Good Guys, (unlike those other VERY EVIL guys), doing all of this responsibly by building Good AI and by communicating these risks to everyone while we build Very Good AI (unlike the ver... (read more)