Our modern society has achieved an amazing degree of division of labour in knowledge and hence specialisation, particularly in science, but co-ordinating that expert knowledge to make it available to society in general is surprisingly difficult. Consider the problems faced by the non-expert in accessing and employing expert knowledge to address particular problems. The politician who wants to know if GM crops are safe; the fisherfolk trying to work out what's happening to all the fish; the parents trying to assess risks of particular vaccinations for their child; and so on.
Note that this problem goes beyond information asymmetry as analysed in economics (although some strategic issues may be relevant) since that assumes that someone already has the right answer to your problem. Instead the uncertainty faced by the non-expert inquirer is more fundamental since the answer to your question may not yet exist. Answering it therefore requires understanding how the world works by understanding how knowledge works. Unfortunately the same skills and knowledge required to see whether one has the right (or at least a good) answer to a problem are the same ones required for arriving at a good answer in the first place. So if you start off ignorant but think you are wise, then you will likely get bad answers but you won't realise it.
1. One must know where to look. Knowledge is now so specialised that a practical problem often falls across multiple domains, with different institutions, people, methods, and (limited) answers. Knowledge therefore has a geography - a spatial distribution of activity and resources.Vaccination safety for example involves medical researchers who produce and test vaccines, but also epidemiologists and medical practitioners who all have different concerns, none of which exactly fit with the urgent question parents have about the safety of their children (vaccine scares have been frequent from the beginning, although, strangely, they rarely cross political borders). Even if one can find out where the relevant knowledge is held, how can one bring it together successfully?
2. Who is an expert (or at least, has relevant expertise)? Within each area different people claim expertise - how is the non-expert to judge who to listen to? Several social psychological phenomena apply here. First, ignorance is a great equaliser - when confronted with an unknown subject area people tend to grant every articulated opinion tentative equality, so even though 95% of relevantly trained scientists may back 1 position (e.g. on climate change), the credibility of alternative ideas isn't affected by their tiny support. Second, ignorance is biased by what one already thinks one knows and how one expects to be persuaded. Non-experts cannot assess the science itself, or even the reputation mechanisms that operate in any particular subfield, so they must do their best to judge who the real experts are based on the intelligibility of what they say and its coherence with how they think the world works. Global climate change, for example, can be explained convincingly as a scientific conspiracy if that is how one already thinks the world works. Note here that most really specialised experts spend most of their time talking to each other in esoteric impenetrable jargon on specific technical questions (in specialist journals that most people can't even access), and have little time or skill for explaining their subject at the basic level required by hoi polloi. Those who are better at public rhetoric therefore have an advantage in public credibility whatever their scientific credentials.
3. Non-experts make an implicit choice about what they want to find out that follows from their characterisation of the problem. There is the geographical range of expert opinion they could focus on, and within that the range of professed experts, and the fact that one's own question never fits neatly into any of the pre-existing knowledge boxes. All of this means that people have to make a real choice about how to proceed that is underdetermined by the structure of specialised knowledge available. Since that choice is rarely acknowledged, the human talent for post-hoc rationalisation has free play. One may believe that one is being led by what the real expert knowledge says, but one is also being led by one's prejudices (for example that GM companies are evil exploitative capitalists, that GM tech must be dangerous because it is messing with nature, etc) in choosing exactly the 'expertise' that suits rather than tests one's prejudices. Of course these biases can include self-interest, although, as in the Japanese (or EU) fishers who want to continue fishing apparently almost extinct species, that self-interest need not actually be rational.
What can one do about these where, who, and what problems? Clearly they are irresolvable insofar as they are characteristic of knowledge itself, but the worst effects can be partly mitigated by better institutions. On the supply side there should be credible public institutions (i.e. with reputations for integrity and reliability) that try to stitch together diverse specialised knowledge boxes, particularly to address important and relatively standard public interest questions (such as drugs policy), and also attempt to present their results credibly to the general public in ways that the public can relate to and believe in. The IPCC is one attempt at this - it could do better. On the demand side, the non-expert 'public' depends on common sense for its assessment of expert knowledge. The more sophisticated that common sense, the more the ordinary inquirer will be able to understand the scope and limits of different kinds of credible scientific institutions, cope with the counter-intuitive results produced by real science (e.g. correlation is not causation), distinguish quacks from experts, and focus on asking the right questions of the scientific evidence rather than seeking the right answers. In addition, wisdom itself requires personal virtues of patience, practical reasoning, careful attention, humility and so on in order to bring an iterative approach to problem solving. That means that the specification of a problem - and one's understanding of the resources appropriate to resolving it - evolves in response to one's assimilation of the answers one achieves to earlier, more primitive questions.