Do there need to be laws for the internet like there are laws in real life? The two realms are very different, so attempts to directly translate real-life rules to the internet are bound to fail.
For example, most laws in real life are to protect A from bad things that B might want to do to him (such as murder or theft). But all such forms of law are unnecessary on the internet because all that B can do to A is send packets to him. A is under no obligation to do anything with these packets - if he doesn't like them he can just ignore them.
So most of the laws that apply on the internet are the more nebulous sort that protect some third party C from B sending packets to A (for example copyright, C being the copyright holder) - censorship, essentially. But since C usually can't even tell when A and B are communicating, such laws are rather difficult to enforce and there is something rather troubling about having unenforceable laws on the books.
Another sort of law is that which protects commons (like a village square) from having houses built on it, or having rubbish dumped on it or otherwise being made unusable for its intended purpose.
Commons as such don't exist on the internet in quite the same way - for every computer on the internet, somebody has to pay for the electricity to keep it running if nothing else. There are "intellectual commons" - protocols and standards which we use to communicate. They can't really be stolen but they can be polluted in some sense - email is quite heavily polluted by spam now, for example.
Combating internet pollution does appear on the face of it to be something that requires laws (and indeed laws have been passed that try to do something about spam - the state of my junk mail folder suggests that haven't been very effective, though). But there is a technical way to deal with such things, and that is through the use of trust networks. The Sovereign Computing guys have some interesting ideas about trust networks. It does seem to me that a trust network layer built on top of the internet and which could provide services to applications like email would be a very useful thing.
Some social network sites are quite popular but these tend to be walled gardens, lacking the distributedness that would be needed to create real trust networks. They all also seem to have little if any concept of transitivity of trust meaning that they are little more than glorified lists of friends as far as their users are concerned.