One objection to these studies is that the workers are looking at artifacts. The so-called transcripts are just noise from accidental transcription. This ties in with the idea that the EST database is full of examples of "transcripts" that don't make any biological sense.
There's another possibility. The regions of junk DNA could be transcribed regularly but the transcripts are rapidly degraded. They do not have a biological function. They are junk RNA.
Arthur Hunt has just posted an article on Panda's Thinb that supports this idea [Junk to the second power]. He describes the work of Wyers et al. (2005) in yeast cells. They show that there is a large class of junk RNA. The take-home lesson here is that you can't assume that some region of genomic DNA is functional just because it's transcribed. It's a lesson that many people need to keep in mind.
Wyers, F., Rougemaille, M., Badis, G., Rousselle, J.C., Dufour, M.E., Boulay, J., Régnault, B., Devaux, F., Namane, A., Séraphin, B., Libri, D. and Jacquier A. (2005) Cryptic pol II transcripts are degraded by a nuclear quality control pathway involving a new poly(A) polymerase. Cell 121:725-37.