Author Topic: 23andme  (Read 2077 times)

Offline Gadget

  • RootsChat Marquessate
  • *******
  • Posts: 57,835
    • View Profile
Re: 23andme
« Reply #9 on: Wednesday 25 October 23 22:47 BST (UK) »
What would they do with such information?

Add - a Google search throws up lots of info:

https://tinyurl.com/mush7d6x
Census &  BMD information Crown Copyright www.nationalarchives.gov.uk and GROS - www.scotlandspeople.gov.uk

***Restorers - Please do not use my restores without my permission. Thanks***

https://www.rootschat.com/forum/index.php?topic=877762.0

Offline emeraldcity

  • RootsChat Member
  • ***
  • Posts: 144
  • Census information Crown Copyright, from www.nationalarchives.gov.uk
    • View Profile
Re: 23andme
« Reply #10 on: Thursday 26 October 23 00:18 BST (UK) »
What would they do with such information?

Add - a Google search throws up lots of info:

https://tinyurl.com/mush7d6x

When it comes to the raw genetic data that's breached we're entering uncharted territory. It's disconcerting to say the least though. The initial breach was specifically marketed on hacker forums as being the genetic info of Jewish people so there's an implied threat there already. The most recent one involving British users promises that it features identifying information on celebrities and European royalty. The rise of sophisticated AI models that can process this vast amount of data is also quite scary.

The secondary kind of data breach involving just information on the relatives profile page is not quite as bad, although it could still potentially be used to identify people who otherwise wish not to be identified.

I'm not happy with 23andme's communication on this. In my case I had no issues with my account security and was affected purely because someone else was - this is a really bad design for a website that handles personal genetic data and should never have been allowed to happen.

I'm also not totally convinced that 23andme have a handle on the scale of this and how much DNA has actually been breached. It wouldn't surprise me if it turns out that hackers found an exploit to also access the genetic data of relatives connected to password compromised accounts.

This is a screenshot of a post by the hacker explaining his motives - it's a bit techy but worth reading: https://www.reddit.com/media?url=https%3A%2F%2Fi.redd.it%2Fphscpvhgbevb1.png

Offline ikas

  • RootsChat Senior
  • ****
  • Posts: 328
  • Census information Crown Copyright, from www.nationalarchives.gov.uk
    • View Profile
Re: 23andme
« Reply #11 on: Thursday 26 October 23 13:10 BST (UK) »
This is a screenshot of a post by the hacker explaining his motives - it's a bit techy but worth reading: https://www.reddit.com/media?url=https%3A%2F%2Fi.redd.it%2Fphscpvhgbevb1.png

Any chance of a translation? How did the hacker get the data? Obviously what the hacker is claiming contradicts statement from 23andMe.

Offline phil57

  • RootsChat Veteran
  • *****
  • Posts: 649
  • Census information Crown Copyright, from www.nationalarchives.gov.uk
    • View Profile
Re: 23andme
« Reply #12 on: Thursday 26 October 23 14:21 BST (UK) »
Any chance of a translation? How did the hacker get the data? Obviously what the hacker is claiming contradicts statement from 23andMe.

I am by no means an expert on the technicalities of all of this, but my basic understanding is this:

The initial breach is claimed to have been made by data stuffing. This is where usernames and associated passwords stolen from another account with a different company (i.e. from a different data breach) have been speculatively tried against 23andMe logins.

Unfortunately it is quite common for some people to use the same username and password across many accounts with different organisations, and datasets of such stolen information can readily be purchased if you know where to look, e.g. on the dark web. So whilst not all the logins tried will have been successful, there were probably quite a few logins achieved, giving access to those users accounts.

The hacker then probably used packet sniffing, a form of capturing the underlying internet data traffic involved in sending and receiving requests such as obtaining the details of DNA matches related to the compromised accounts, and identified a sequence or sequences of data strings that could be sent in a particular way to access the information for any individual they wanted - as I read it from one of the posts on the subject, within an average of 50 attempts at sending particular data strings for each account they wished to access. It seems that the data required to be sent to access those accounts might have followed certain patterns that the hacker was able to identify and use to their advantage, and was not as random as 23andMe perhaps believed when the software was designed.

Of course to do this manually would take hours for each individual attempt, and probably more than a lifetime to obtain data on millions of users, but once the initial data strings in the transmitted packets and the patterns in the sequences generated were discovered, an automated computer script could be created that would generate multiple such request in a fraction of a second and capture the information desired very quickly.

I gather that 23andMe have only admitted to the data stuffing attack using stolen username and password credentials, but the rest as claimed by the alleged hacker elsewhere seems plausible in view of the fact that 23andMe have since disabled their DNA tools and now My Heritage have also done likewise, possibly as a precaution until they can investigate and verify the security of their own system, or perhaps because they now know that it is also vulnerable to a similar technique.

I may be wrong, so don't take the above as being factually correct, but it's just my interpretation from what I have seen in various posts on the subject elsewhere.
Stokes - London and Essex
Hodges - Somerset
Murden - Notts
Humphries/Humphreys from Montgomeryshire


Offline ikas

  • RootsChat Senior
  • ****
  • Posts: 328
  • Census information Crown Copyright, from www.nationalarchives.gov.uk
    • View Profile
Re: 23andme
« Reply #13 on: Thursday 26 October 23 14:46 BST (UK) »
Thanks Phil for your explanation. I assumed all data exchanges between user and 23andme website would be encrypted. If they were using packet sniffing I presume the data exchanges were not encrypted?

Offline phil57

  • RootsChat Veteran
  • *****
  • Posts: 649
  • Census information Crown Copyright, from www.nationalarchives.gov.uk
    • View Profile
Re: 23andme
« Reply #14 on: Thursday 26 October 23 16:34 BST (UK) »
I don't know the answer to that. I'm not a customer. But even if that were the case, bearing in mind this may be a very simplistic explanation. Suppose as seems to be inferred, you have a possibly large number of usernames and passwords in plain text that you now know work on 23andMe. You know that to login, you have to send those two pieces of information. By examining the data packets going back and forth, even if they are encrypted, you know that certain strings will be constant to every login attempt, whereas others will contain the encrypted username and password. If you know what the plain text content of any parts of those conversations should be, and/or you are able to identify certain parts as containing the username or password, then as you know the plain text phrases, and with multiple usernames and passwords you've likely got the whole alphabet and range of numeric characters covered, probably several times over. With enough cross referencing of the communication data, it should be possible to crack the encryption. There may be other security features involved, such as random seeds or tokens normally specific to each user's computer. But as you are entering all the details for each account from a single computer, again it may be possible to identify that or work out the algorithms involved.
Stokes - London and Essex
Hodges - Somerset
Murden - Notts
Humphries/Humphreys from Montgomeryshire

Offline Gadget

  • RootsChat Marquessate
  • *******
  • Posts: 57,835
    • View Profile
Re: 23andme
« Reply #15 on: Thursday 26 October 23 20:50 BST (UK) »
I think I'm going to ask for my account/data to be deleted - note that you have to ask  them ::)
I've  already got most of my info downloaded.

Also, I'll keep my eye on what is happening to My Heritage.

Will Ancestry be next??

Gadget



Census &  BMD information Crown Copyright www.nationalarchives.gov.uk and GROS - www.scotlandspeople.gov.uk

***Restorers - Please do not use my restores without my permission. Thanks***

https://www.rootschat.com/forum/index.php?topic=877762.0

Offline phil57

  • RootsChat Veteran
  • *****
  • Posts: 649
  • Census information Crown Copyright, from www.nationalarchives.gov.uk
    • View Profile
Re: 23andme
« Reply #16 on: Friday 27 October 23 11:28 BST (UK) »
More info and advice re the 23andMe issue on Roberta Estes home page here:

https://dna-explained.com/
Stokes - London and Essex
Hodges - Somerset
Murden - Notts
Humphries/Humphreys from Montgomeryshire

Offline Gadget

  • RootsChat Marquessate
  • *******
  • Posts: 57,835
    • View Profile
Re: 23andme
« Reply #17 on: Friday 27 October 23 12:08 BST (UK) »
Thanks for the extra link, Phil.

I  was only interested in the health aspects and maternal haplogroup that they provided.

I have requested deletion as the site seems so vulnerable to a breach.


Gadget

PS - I'd already opted out of DNA Relatives and have always used a unique password.
Census &  BMD information Crown Copyright www.nationalarchives.gov.uk and GROS - www.scotlandspeople.gov.uk

***Restorers - Please do not use my restores without my permission. Thanks***

https://www.rootschat.com/forum/index.php?topic=877762.0