I feel like backend is more predictable. Known inputs and outputs (think an API). I can get ChatGPT to act like an API with a single query; but actually building a front-end application not so much.
But also only for sort of trivial applications. I run into real problems all the time. It's true that a lot of applications are trivial and often there is no budget for the interesting parts - and that's about to change. So perhaps this development is a good thing.
I wonder how you are currently predicting a random api change, login flow change, business reorganisation change, inheriting and integrating a new microservice to the ecosystem of n existing other applications.
What do you imagine the backend is doing? Just providing a connection to a database somewhere?