That's genius. You can just casually drop at an interview, "And, of course, I've been published in Nature and Science," and sound like a total BAMF. What a brilliant scam.

What you've described is certainly the model that is used at many schools, however it's not applicable everywhere. I've attended several schools with strong honor codes, and which allow students to self-schedule their exams within certain windows of time. At my current professional school, for example, we typically have a three day window in which to take an exam. We can show up at one of several exam locations at any point in time, day or night, within that window, sit down, take it, and leave. This is a very convenient system, and I greatly appreciate the flexibility that comes with it. However, when I walk out of the exam room, the exam most certainly does not become public information. This system absolutely depends on the rules of the exam extending outside of the exam room. If I could talk to one of the students who took the exam earlier than me about what was on it, that would absolutely be cheating.

I was going to post the exact same things. If getting information about previous exams is cheating, it's cheating. The people designing a test get to define the parameters of what's cheating and what's not. On some tests you can use a calculator. Some tests are open book. For others tests you aren't allowed to look at previous administrations.

Now, the question of what's legal is a separate issue. You can cheat on an exam without falling afoul of the law. Depending on the exam, you could follow the rules of the exam while breaking the law. The issue of cheating is exactly orthogonal to any questions of legality (excepting possibly exams with legal consequences - so your Step I and II exams might be a different story).

People have long assumed that vaccinia was cowpox, but modern biotechnology makes it clear that they are two different viruses. Intriguingly, no one knows where vaccinia came from. It's a total mystery.