For me, one idea that is dangerous and possibly true is an extreme form of evolutionary psychology — the view that many behaviors of modern humans were genetically hard-wired (or soft-wired) in our distant ancestors by natural selection.
The reason I say that this idea might be true is that we cannot be sure of the genetic and evolutionary underpinnings of most human behaviors. It is difficult or impossible to test many of the conjectures of evolutionary psychology. Thus, we can say only that behaviors such as the sexual predilections of men versus women, and the extreme competitiveness of males, are consistent with evolutionary psychology.
But consistency arguments have two problems. First, they are not hard scientific proof. Are we satisfied that sonnets are phallic extensions simply because some male poets might have used them to lure females? Such arguments fail to meet the normal standards of scientific evidence.
Second, as is well known, one can make consistency arguments for virtually every human behavior. Given the possibilities of kin selection (natural selection for behaviors that do no good for to their performers but are advantageous to their relatives) and reciprocal altruism, and our ignorance of the environments of our ancestors, there is no trait beyond evolutionary explanation. Indeed, there are claims for the evolutionary origin of even manifestly maladaptive behaviors, such as homosexuality, priestly celibacy, and extreme forms of altruism (e.g., self-sacrifice during wartime). But surely we cannot consider it scientifically proven that genes for homosexuality are maintained in human populations by kin selection. This remains possible but undemonstrated.
Nevertheless, much of human behavior does seem to conform to Darwinian expectations. Males are promiscuous and females coy. We treat our relatives better than we do other people. The problem is where to draw the line between those behaviors that are so obviously adaptive that no one doubts their genesis (e.g. sleeping and eating), those which are probably but not as obviously adaptive (e.g., human sexual behavior and our fondness for fats and sweets) and those whose adaptive basis is highly speculative (e.g., the origin of art and our love of the outdoors).
Although I have been highly critical of evolutionary psychology, I have not done so from political motives, nor do I think that the discipline is in principle misguided. Rather, I have been critical because evolutionary psychologists seem unwilling to draw lines between what can be taken as demonstrated and what remains speculative, making the discipline more of a faith than a science. This lack of rigor endangers the reputation of all of evolutionary biology, making our endeavors seem to be merely the concoction of ingenious stories. If we are truly to understand human nature, and use this knowledge constructively, we must distinguish the probably true from the possibly true.
So, why do I see evolutionary psychology as dangerous? I think it is because I am afraid to see myself and my fellow humans as mere marionettes dancing on genetic strings. I would like to think that we have immense freedom to better ourselves as individuals and to create a just and egalitarian society. Granted, genetics is not destiny, but neither are we completely free of our evolutionary baggage. Might genetics really hold a leash on our capacity to change? If so, then some claims of evolutionary psychology give us convenient but dangerous excuses for behaviors that seem unacceptable. It is all too easy, for example, for philandering males to excuse their behavior as evolutionarily justified. Evolutionary psychologists argue that it is possible to overcome our evolutionary heritage. But what if it is not so easy to take the Dawkinsian road and "rebel against the tyranny of the selfish replicators"?