This Friday and Saturday, I took part in a working group meeting of 30 open government advocates, organized by Carl Malamud and Tim O'Reilly, to develop a set of Open Government Data Principles.
One of the few bright spots in United States copyright law has always been that data produced by the government is, in theory, in the public domain. While there have of course been encroachments on this doctrine from time to time, it has generally been been held to in practice as well as in theory.
Unfortunately, being in the public domain isn't necessarily the same as being online and accessible in reasonable formats via modern protocols. For example, Carl Malamud has spent a fair amount of effort prying the raw records of copyright registrations out of the U.S. Copyright Office at the Library of Congress and putting them online in a much more useful way than the government ever had. Similar stories abound among those with experience extracting electronic data from governments.