OpenAI accidentally deleted potential evidence in NY Times copyright lawsuit
In a court filing, lawyers for The NY Times and Daily news say that OpenAI accidentally deleted potential evidence against it.
You're viewing a single thread.
In Spain, in a major political corruption trial, a party turned in as evidence some drives that had been erased by Dban 7 times. They argued that it was routine to do seven passes.
It is... It's literally a preconfigured option on the dban selection list.
Source: My memory... but if that's not good enough, here's wiki too.
https://en.wikipedia.org/wiki/Darik%27s_Boot_and_Nuke
and DOD 5220.22-M (7 passes) are also included as options to handle data remanence.
It's an option, but not the default. It takes forever to run, so someone using it is being very intentional.
It's also considered wildly overkill, especially with modern drives and their data density. Even a single pass of zeros, the fastest and default dban option, wipe data at a level that you would need a nation state actor to even try to recover data.
so someone using it is being very intentional.
Not if you're used to taking DoD requests. It was my default for a very long time because I simply defaulted to it for compliance reasons.
It’s also considered wildly overkill
Absolutely is. Doesn't mean that people like me aren't out there in droves.
But SSDs make this all moot and HDD are being phased out of many environments. SSDs with chucking the key is more than sufficient as well.
Congrats? DBAN was made prior to 2006... IT people existed before 2006. What's your point? You think that people just spawned into existence in 2006 with decades of IT knowledge? So like I said... "It WAS my default for a very long time because I simply defaulted to it for COMPLIANCE reasons"... eg. my contracts at the time required it and I ran boatloads of wipes.
Regardless... DOD 5220.22-M now states
The National Industrial Security Program Operating Manual (NISPOM) is now Part 117 of Title 32, Code of Federal Regulations.
So let's go look at the NISPOM stuff which says... NOTHING! So what you end up with is companies referencing the old DOD 5220.22-M because old government contracts will actually say that specific document in contracts as something that must be adhered to for a long long time. So even though it "died" on 2006, contracts may not be renewed for some time after that which still keeps the document alive.
Now DOD 5220.22-M actually specified and defines short wipes (3 pass) and long wipes (7 pass). And in theory, could be superceded by NIST 800-88 (and probably is the default on modern contracts). And regardless of all of that... DoD internally has it's own standards, which after wipe often requires degaussing or outright destruction of the disk, I remember having a dedicated device for it that would document serials and stuff. I'd have to pull up my army documents to remember which specific rules required that type of stuff, but I'm not going to dig out shit from 2010 just to argue with someone on lemmy.
So I guess this boils down to... The world didn't spawn into existence in 2006. People are older than 2006 and are allowed to talk about their experiences from before the "old times".
Edit: And in current contracts... all our shit is NVMe and secure erase. But I'm willing to bet muscle memory would still kick in for me if I saw the DBAN screen.
And honestly, if you're going to do a single pass, might as well do multiple. It doesn't take any more of my time for 1 pass vs 7, assuming I only have a handful to do. I'll probably just start one before I leave for the day, swap to another when I come in, and repeat until the pile is cleared.
If something is worth doing, and overdoing doesn't take any extra effort, I'll overdo it.
That was basically the workflow. On smaller drives you could do one when you get in, one at lunch and one before you left. Eventually drives got large enough that it was just once in the morning and once before leaving.
I’ll overdo it.
Half the contracts you didn't know if they wanted the short wipes or long wipes. So you just do long wipes to cover your ass. It's not like there was a rush, it was a simply menial task that became a second nature set of bashing the keyboard. Like typing some of my passwords and pins... I have no fucking clue what they are anymore... but put in front of the keyboard and I can type them by muscle memory.
Yup.
I didn't do this all that much for a job, but I did work in IT tech support for a couple years and helped out w/ some decommissioning in a dev role, and I would usually put it on a higher setting. Why? Idk, why not? Best case scenario, I prevent a scandal from some really motivated attacker, worst case, I check on it and it's not quite done, so I check on it later.
I'm discussing this comment :
https://sopuli.xyz/comment/13141026
the one that you initially replied to talking about recent Spanish court case where the defendants used a 7x wipe on some drives that were required to be retained as evidence.
Im well aware sysadmins existed before 2006, and also don't see how that's relevant in context. Security practices change over the course of 18 years in IT, as they have for secure wiping data.
So am I. I'm not sure what you think wasn't relevant. It's a literal DoD spec. Yes that spec is outdated, but it's still in Dban.
You coming out of nowhere talking about how the DoD spec itself is "dead" doesn't change the fact that it's available and probably still used by many people out there. I'm willing to be that several companies have the old DoD spec embedded in their own SOPs. And I was always talking in the context of the contract work I did long ago which WAS to the old DoD spec regardless.
Okay so what you think is wildly overkill, is about 10% of the effort some organizations go through to make sure data is not restoreable.
My org shreds discs entirely with a mechanical grinder, so I'm well aware of overkill.
Multiple overwrites being unnecessary isnt really an opinion. Here is the company that owns dban agreeing with security orgs like NIST, that anything past 1 write is unnecessary. .
I think the issue comes down to whether the org in question does that 7 passes consistently on all discs, or if it just so happened to start that policy with those that had evidence on them.
I think the issue comes down to whether the org in question does that 7 passes consistently on all discs, or if it just so happened to start that policy with those that had evidence on them.
No? If 1 is sufficient, any additional shouldn't matter in any considerations at all. Could have simply been somebody who hit the preset on accident.
Were they erased when the investigation started or was it done time before?
