I would agree on the value of such assessments if the only thing anyone wanted to do was use it in as ham-fisted a way as you suggested. Using it instead to understand the opposition and then find more productive ways to reach them would be much more useful.
And if we're stupid enough to judge climate science by the success rate of weather predictions(!?) or the morons who helped create the financial meltdown, then call in the dogs, put out the fire, the party's over.
As for humility--yes, that's definitely in order. I was a professional programmer for a long time, and I have the battle scars to show how spectacularly systems can fail due to arrogant designers and programmers. But how much evidence do we need before we say, "OK, we might be off by +/-10%, but the probability of Bad Stuff Happening for this particular reason (CO2 emissions) is so high we have to take action, because the expected value of being wrong is so much less than the expected value of being right and not taking action"?