This is nothing but appeasement. Publishers at large won't know what hit them once searchgpt drops. The written internet is doomed.
As someone who used searchgpt and is also building a competing product, I can say this is simply not going to be the case. LLM assisted search leaves just too much to be desired and there is no future in which this is the primary way for humans to consume information. It needs to be augmented with credible sources of information or rather using LLMs only makes sense once you have access to credible source information in the first place, and want to have an augmented version of the information you consume (and are happy with non discrete outcomes).
A multi-billion dollar cheque? Voluntarily relinquished or via a settlement?
In that scenario, in a few years it will be out of date and doomed too.
It will train on what people say in their cellphones or in front of their Smart TVs or in their deeply connected cars. If data is being transmitted after encryption (within a closed and undocumented chip) to a server in a country that doesn't cooperate with authorities, then the same data is pseudo-anonymized and bought back on a different channel, that will make very hard to stop it.
And the video internet is under ever-increasing enshittification.-
For CN Traveler, I'm pretty sure that readers want an actual picture of the interior of the new hotel, not an AI rendering of something that as yet has few actual pictures to train the model on.
(For that matter, Vogue might have that same thing going on with fashion.)