A couple of points:
1. What does it mean to say that humans are terrible objective observers? In some ways we seem to be extremely accurate objective observers. We make excellent predictions about weather, we can build complex industrial instruments; day to day our understanding of the world is almost wholly accurate. This is likely why, evolutionarily, we accept our perceptions of the world as true unless we have strong evidence to the contrary (it helps people survive when they don’t think that a lion attempting to kill them is a hallucination). So in what way are people terrible observers? Only relative to some immaculate perceiving machine, whatever that would mean. So yes, it appears that people can be tricked and harassed into false beliefs. And yes, people have some systematic biases like overconfidence, hindsight bias, and, as George mentioned, a kind of herd bias. But the real story here is how good we are at perceiving the world, not how bad.
2. Related to this fact is that because people start off with the assumption that they act somewhat rationally and correctly assess most situations, psychology will naturally focus on those areas where we’re weakest. This is probably a good thing, but at the same time the fact that the New York Times never publishes a story like “Human eyesight incredibly good at detecting things it needs to survive, and is astonishingly reliable at relaying detailed visual information to the brain” means that there’s a publication bias not only to look into human shortcomings but to publish more frequently when they’re found than when they’re not. And even within those that are published, people love to focus on how bad humans are at reasoning. For example, studies putting people’s statistical reasoning in a bad light are cited more than 5 times as frequently as those that show us to good effect.
3. There are also a couple more general problems with the research George cites. One is the sample populations, which are almost universally atypical of people worldwide, consisting of Western, educated, industrialized individuals studying psychology. The small sample size of these studies (35 in the experiment George mentioned) is also a problem, although the Milgram study does seem to have held up to replication.
4. More to George’s concerns about juries, there’s a pretty obvious difference in that there aren’t planted actors being paid to lead the jurors by their noses. But more to the point, what’s the alternative? Maybe judges are to be preferred to juries, but conformation bias doesn’t seem to the primary problem with juries. As far as the conformation bias does appear in the courts, it appears to work in the opposite direction on judicial panels, where the ideological minority in three-judge panels often prevail over their colleagues; for example, controlling for ideology, the presence of a female judge on a panel strongly increases the chance that the panel will rule in favor of the plaintiff in employment discrimination cases.