The library of Congress, along with others, has carried out a study on the impact of copyright on digital preservation
link. The 200 odd page report comes to the conclusion that 1) laws were enacted in an analog era, thus do not understand digital preservation; 2) there is little in the sense of mandatory preservation of digital works, like there is for paper copies at the Library of Congress; 3) archives have resorted to piecemeal agreements to digitize works instead of having blanket approval, due to copyright laws.
The report recommends in particular that the law should define "preservation institutions" that can digitize at risk copyrighted material, that they should be allowed to preserve in various formats, that they should be allowed to harvest material pro-actively, irrespectively of the medium and the type of material. In short, the preservation institutions should be granted privileges that allow them to bypass copyrights in a wholesale manner, for the sake of preservation.
Hat tip: DigitalKoans.