Regular citizens asking elites tough questions usually fails (HuffPost example)
Eliot Spitzer has posted some questions for bankers, see that link for the details and the link. And, probably in an attempt at just ginning up page views, the Huffington Post has invited their readers to submit questions. The submissions once again illustrate how regular citizens asking the elites questions serves the interests of those elites. The regular citizens think they're taking on the elites, when in fact they help them. That certainly sounds elitist in itself, but it's actually just a reflection of a reality that's happened time and again; see the popular voting systems for several past examples.
That doesn't mean that regular citizens shouldn't have input or be able to ask questions if they can come up with good ones. It just means that any selection process (see the link above) needs to be designed so that weak questions don't crowd out those that would have a far greater impact. In some cases, the selection process seems to have been designed to do exactly that: politicians don't like being asked tough questions and they know what's going to happen when they allow regular citizens to vote on the "best" questions: the easily-answered puffballs are going to rise to the top.
In the current case, the first question the HuffPost is promoting is simply "Mortgage loans modification", which isn't a question.
The second is "What kinds of internal controls and fraud training were and are in place?" Imagine yourself being asked that question: wouldn't you say "we had extensive internal controls and fraud training in place" and then simply segue into reading from your company's brochure?
And, the third is "Why shouldn't the big banks be broken up?" Imagine the brochure-speak about economies of scale and on and on that would generate.
That doesn't mean they're all bad, but it does mean that a) those who don't have enough knowledge and experience to ask tough questions should put their egos aside and let those who do ask the questions, and b) systems which actually intend to generate tough questions (rather than setups) should be designed so the best questions rise to the top.