"show the dangers of blanket trust"
Ignores a fundememtal problem of trust or the limitation of it, there is no point at which trust cannot be abused.
The first area is that people have to be able to perform their jobs, implicit to this is that they have to be trusted with confidential information to do this (ie a receptionist needs the company telephone directory to connect calls, but they should not give the information out to callers etc).
The follow on from this is that at some point in the organisation somebody has to be trusted sufficiently to be able to see all information (even though they may not be aware of it's existance untill they need to see it). Even in the most secret of organisations this is true.
However the cost of trying to limit access to more information than is actually required is usually prohibitivly high so only certain types of organisation do this (and usually not very well).
These costs are not just in the systems put into place to limit trust but also in the form of lost oportunities where trust has prevented the linking of information to provide critical information to the organisation.
So you end up with a trade off, the cost of trust vis the cost os secrecy. As your first problem, how ever there are other problems as the US goverment amongst otheres have realised there are two basic types of secret that have to be protected,
1, Those that are known to people
in positions of trust.
2, Those that can be deduced by
those who are not trusted, from
"publicly available" information.
The "agrigation of information" is a very very difficult problem (see Ross J Anderson's book ( http://www.cl.cam.ac.uk/~rja14/book.html ) or his home page ( http://www.cl.cam.ac.uk/users/rja14/ ) for more information on this problem.
In the UK there have been a number of cases where the "agrigation of information" has given rise to attempts at prosecution under the Official Secrets Act. In one case a journalist was accused of releasing secret information (The address of GCHQ) and much to the prosecutions embarisment he simply produced a copy of an internationaly circulated magazine (Wireless World) that had a full page job advert for GCHQ with the address prominently displayed (it was shortly after this the case colapsed).
Then there are problems with releasing data to people who need it in an anoynomous way (such as medical records to drug researchers). Inveriably it is not possible to stop inadvertant leaks of information, especially when you only control or have access to part of the available data set.
There are a whole load more problems to do with "Trust" it is an extreamly difficult subject, and has been the subject of much research over the years and in the case of computer security well over 40 years in the public domain.
Basically there is no level of trust that in some way cannot be abused either accidently or deliberatly, most organisations are aware of this, and have to accept the fact that every so often there is going to be a "bad apple in the barrel".