"97 percent of climate scientists believe human activities are causing global warming."
That's not a scientific statement, it's a political one.
Wrong, it is a scientific fact. Some scientist measured a trait (belief that human activities cause global warming) in a population (climate scientists), and reported the number in said population with the trait. The correlation between number of scientists that believe in a theory and the predictive accuracy of said theory is a separate (and also scientific) question, and also part of the very useful rhetorical device known as "argument from authority".
It is, however, incredibly easy to insert politics into polls, by even minor rewording of the poll question vs the report, or by one of the many possible methods of biasing the sample population. For example, for the question "Do you believe that human CO2 emissions cause an increase in temperature?" 100% of the people who respond "no" are ignorant of some very basic science*, or lying. Whereas if you ask the question as "Do you believe, as a matter of scientific fact, that if it were not for human activities, half or more of the increase in average global temperature would not have occurred?", an answer of "yes" indicates a high level of confidence in an analysis requiring reams of data and hours of supercomputer time. Yet either of those questions could (and would) be translated to "percent of [tested population] that believes in AGW".
*CO2 is a greenhouse gas, which means it is more transparent to energy radiated from the sun than from the earth. Note that the wording of that question avoided all of the things, such as "how much" and "for how long", that would make the answer relevant to anyone, except as a test, or as propaganda.
Science doesn't vote, it either provably is or it isn't.
Nope, it either makes good predictions or it doesn't -- truth is for the mathematicians and philosophers. Much of science has been provably false, eg Newtonian physics and various specific theories of evolution, and those theories have been discarded and replaced with different, slightly better approximations. That the refinements often retain much of the original theory and sometimes the same name doesn't mean the original was not provably of inferior predictive power (not provably false, because it is about predictive power not about truth). Note that it is the predictive power of science (not the truth) that is useful for designing technology. For example, even if Newtonian physics is false, people still use it instead of relativistic quantum mechanics, if they can get away with it.
Sorry to be such a stickler for detail; it's just that in my experience whenever someone starts talking about "truth" in the context of science, instead of "accuracy" or "predictive power", it's generally an indication that I'm about to be treated to a load of rhetoric.