Climate 'scientists' might confidently state that the world will warm by X.YZ degrees in the next 20 years. What they don't ever tell you is the uncertainty in their prediction, basically because it is nearly impossible to quantify.
Take a look at these climate model results:
http://www.ig.utexas.edu/people/staff/charles/uncertainties_in_model_predictio.htm
Which model is closest to correct? Each model makes large numbers of different assumptions about how to mimic radiation, atmospheric turbulence, adequate atmospheric and terrain resolution, and any number of other phenomena. As the actual system is highly nonlinear, even a small uncertainty in the initial conditions can lead to wildly different results even if we had all of the equations exactly correct (which we don't, most are modeling approximations to make the problem tractable).
The best that can be said is that it seems probable that the Earth will be getting warmer. The questions are how much and how quickly? Having a number of predicted outcomes means that there is a range of policy decisions, and politicians can cherry pick the outcomes that resonate with their ideology. If a politician seizes on a prediction that indicates that warming isn't a big deal, they will push for the status quo, especially if the are already benefiting from the status quo. Or maybe they will seize the worst case outcome which suggests major societal upheaval is required to remedy it.
Also, as it is a chaotic system, there really is no way to determine if your attempts to control it were even meaningful, even in hindsight, as chaotic systems change non-linearly without human input. That is the argument of the folks who believe that AGW isn't occurring because the world was warming before humans came along. Others think it's all our fault. Without the ability to spin up a human-free Earth 2.0 as a control, it is very difficult to tease apart what is what.