I'll try to be brief.
I agree with you. I think there needs to be accountability, especially around datasets that government is using to make decisions.
I am interested in crowdsourcing, but I think there are incredible limits around how to do it. Even in the example of peer to patent—it's a wonderful example—there are very tight constraints around what it makes it work. It's very easy to use crowdsourcing to disprove things. You may be identifying cases in which something is actually not true, such as identifying patents that are not valid; it's also great to identify datasets that are in error. It's much harder to use it to identify what is actually truthful or is actually a fact.
So one of the nice things is that we should be treating our open data portals as an engagement tool because they're actually a wonderful way to crowdsource errors, not because we want to find errors and make people accountable. There are always going to be errors in the data, so let's surface them more quickly so that we can then get to better quality data faster, so that governments make better decisions with more reliable datasets.