I do large scale computer modeling for a living - I've been in the modeling and simulation realm in one form or another for over 3 decades. Let me be perfectly clear here - these models are not predictive. Say it with me again to make sure you got it - these models are not predictive.
That means they cannot be used to predict actual conditions even 10 years out from present time, let alone 100 years or so. The models simply do not work for that use. There is no rational, defensible argument that can be made that what the model predicts will represent actual conditions. It represents one possible end state out of a large continuum in many dimensions. Any reasonable sensitivity analysis will show that there is very little confidence in any one end state. Very little. When models like this are abused in this way that confidence can quickly drop to tiny fractions of one percent... In other words, whatever it says is going to happen, there's about a 99.xxx% chance it'll be perceptibly, even significantly, something else.
speaking of model predictions, it’s long been mathematically proven that weather will NEVER be accurately predicted beyond 72 hours because the atmosphere, oceansphere, and lithosphere are all chaotic systems. Since “climate” is influenced by all of those as well as solar conditions and interior Earth conditions, then what do you suppose the outer boundary is for accurately predicting “climate change”?