I tried to enter an answer to the concerned professor (Slashdotted) who wondered what to do about Internet-enabled cheating, but you have to register to enter a comment etc., etc.  So here goes:

It is not that hard to solve (or, at least, significantly reduce) cheating. Here is what you do:

1. Have students turn in papers electronically through a service which checks for plagiarism (Personally, I use Blackboard’s SafeAssignment, which works fine, though Blackboard itself is crap). Yes, it will cost the university money. Control mechanisms do. Consider it an investment in academic reputation.
2. Institute a rule that any student submitting a term paper can be subjected to an oral examination about it within a specified time, making paper outsourcing risky. My institution has this in their student handbook.
3. Use multiple methods of evaluation, including class participation. This makes the whole course an evaluation, encourages preparation throughout the course, and might teach you something new.
4. Use fresh examples and/or new and ingenious questions every year, so that the pool of available papers to plagiarize or ready-made Wikipedia entries to amalgamate is reduced.
5. Design the content and teaching of your courses so that they value insight and deliberation rather than repetitive fact checking (for which you should use sit-in exams).

It’s not that hard. It just means structuring the control mechanisms to the content of your course, and getting to know your students well enough that you have a multidimensional view of their abilities.

PS: Incidentally, here is a to-the-point comic (removed, link rot) about this issue. Trouble is, not enough college professors read Wikipedia.

PSPS: The comments to this piece seem to take the same viewpoint, aside from lamenting the fact that teaching college has gone from scholarship to babysitting. It has, and that is lamentable. But if you are going to do babysitting, at least do it well, in a way that does not punish the real students…