I believe that there's a sensitivity threshold for laser gyros at which the Man gets interested. I've been to a colloquium where the speaker noted that foreign graduate students weren't allowed to work on gyros with sensitivities greater than a specific value. They were careful not to optimize beyond that threshold (sensitivity was ancillary to their science measurement).
I've also seen a civilian science project rename themselves in order to differentiate themselves from missile guidance systems. In general though, I don't think it's much of a problem for science. The sudden classification of a promising technology is a strong signal to other states that the technology is of practical use.
In CS terms, the day that quantum computation goes dark is a day you know that RSA is over.