The problem is that reality doesn't always work the way we envision. I'm pretty sure Marx didn't envision Stalin as the most natural outcome of his work. Or Mao. Pol Pot? Um. Yeah.

It's very hard to understand big complex systems with lots of feedback loops. But usually smallish perturbations result in at least someone predictable outcomes.

So while I agree that academia is too cautious, it's also very very easy to (attempt to) overreach especially on big social agenda issues. This doesn't mean we should do nothing, but it does mean that we should test our ideas stepwise, not just throw out huge systems and try to invent new ones from scratch.