The thing is, having your area of interest be all pre-printing-press/pre-widespread-functional-text-literacy/etc always makes posts like these fascinating.
Like my darlings we are putting as much of our information and record keeping and so on in monuments of stone and of acid free paper and so on as anyone used to. Are you kidding me? Our countries are littered with shit with engraved plaques or written details in stone, and equally littered with print copies in archival files or bound into books.
Is there information that exists only in digital form that will be otherwise lost? Yeah, 100%, but it's all information our ancestors never had and wouldn't've been able to store either through its actual nature or because all of that shit was done orally or because it was done on true ephemera (sand tablets, wax tablets, all the fucking tablet forms they used and wiped and used again that weren't a brief era of clay that baked hard when lit on fire.
And there's still so much information from those societies we will never have because of how much never got written down on tablets.
Interestingly, post-printing-press researchers often also overestimate the proportion of information they have vs what actually existed ever, because once you hit the press and mass book binding there are so many records proportionately that it's easy to forget that the vast majority of people interacted with the written word maybe a few times in their life for a long time, and then maybe a few times in a year; then it's easy to forget the mountains of records that were simply outright destroyed because they were no longer of use, and because nobody had anywhere to put thousands of boxes of records, and because they were only ten, twenty, fifty years old, they weren't worth shit yet.
Are we "losing" millions of [pick your measurement here] of data all the time because the technology isn't good enough to preserve it? Yeah, sure.
. . . we always have been.
The threat of data corruption is not a lacuna in history, it's a sudden gap in working, ongoing functional information that we need to keep existing systems working right now, because of format changes and somebody dropping the ball on file transfer. Archaeologists in a thousand years will still have more records than they could even REMOTELY hope to go thru in their lifetimes, but we might be fucked for a continuum of epidemic disease data from the last fifty years, which will become relevant a lot faster than that.