Here's a note on an ugly situation: when a post-doc publishes a paper without the permission of the principal investigator. Now, this is a fairly rare situation, but still not as rare as you might imagine - the article itself has several citations, and it quotes a journal editor who's seen it happen a number of times.
In most of these cases, there seems to be a more fundamental confusion about ownership of data, with publishing as the sequel. People leave a research group with their piles of results, and decide that since it's theirs, that it's time to get it out into the literature with their name on it. But as the article points out, if work is done under NIH funding, then the results belong to the institution, and the grantee/PI is the person who decides when and where things are published. You may, as a grad student or post-doc, feel that the data you worked so hard to generate are rightfully yours, but most of the time that's just not the case.
In industry we have our own disputes, but this isn't one of them. There's rarely any argument about ownership of data: that's all company property, and you sign documents when you're hired that explicitly spell that out. And publication is rarely as bitter a business as it is in academia (where it's the coin of the realm). We argue about whether a particular project is advanced enough (or dead enough, more likely) to be written up for a journal, but these are secondary questions.
Who gets on the patent is a slightly bigger question, but it's not like you get a cut of the profits based on whether your name is on the list. That's as opposed to Germany, where that's exactly what happens, and I've often wondered if we should try that here. That system leads to some elbow-throwing when it comes to inventorship on a hot project, but it also leads to everyone having a clear idea of the legal requirements to be an inventor. Ownership is, naturally, not in dispute at all. Every invention realized at the company is company property, too (those same documents take care of that back when you're hired on).
So while rogue academic publishing is a known phenomenon, rogue industrial patenting isn't. Well, as far as I know it isn't - anyone have an example of someone who tried to get away with it?