I thought this would be more about stylometry but it's mostly about users literally posting the same identifiable information across multiple services, including in one example their age, dog name, profession.
It's all classic dox profiling techniques. Even the things like spelling differences being regional signals and commonality to specific things being discussed.
It's why one has to think about what is being posted to which community if using different identities, rather than posting the same things across all of them. Though any such effort would be a waste if reliant on some non-public info that later was exposed in a database breach which tied together previously unrelated profiles.
I’m curious if an LLM-based defense for this could be made. Like a browser plugin that warns you if you type identifiable information (like occupation) into a text field, and highlights turns of phrase that are “unusual” enough to be identifiable.
There was a tool shared here that could show which accounts belong to the same person based on the writing patterns. Can't remember the name, but it found my old accounts on HN pretty accurately.
Anonymous account unmasking represents a new threat to anonymity.
not just this technique with llms, but the earlier text similarity one.
But I think it would be generally easier to counter in the same way.
Use an llm or heuristics to pose as someone else.
not only do you erase your traces, you add false positives in to the system which reduces the overall effectiveness of these techniques in the future.
A bit of poisoning the well.
I hope eventually an easy to use tool, with maybe a small local llm, can make it easy enough to do this, so that any future deanonymization attacks would be too untrustworthy to rely on
> If you request deletion of your Hacker News account, note that we reserve the right to refuse to (i) delete any of the submissions, favorites, or comments you posted on the Hacker News site
Probably not GDPR-compliant then if comments can be deanonymised by LLMs.
This is probably the worst piece of policy on whole HN. It has a evil feel to it. If HN wasn't so interesting/valueable, this would be the single reason NOT to use it at all.
If they can help to deanonymize you, they must contain something personal.
Writing pattern are pretty personal, certain spelling errors too, or the choose of words.
Absolutely anything relating to an anonymous person could help deanonymization, so that implies that anything relating to any person is personal data. Is that the GDPR’s position?
From ico.org.uk:
“ It is important to note that opinions and inferences are also personal data, maybe special category data, if they directly or indirectly relate to that individual”
From gdpr-info.eu:
“ Subjective information such as opinions, judgements or estimates can be personal data.”
So yes. HN is in violation of the GDPR. I had already filed a complaint about this policy at my local GDPR authority.
Only if said users happen to commit OPSEC failures themselves. LLMs aren't magic...
If someone can figure out who I am or what city I live in just by this username or my comments (with proof), I'll personally send you 500,000 JPY. I'm quite confident that's not going to happen though.
The paper referenced in the article does not even explain their exact testing methodology (such as the tools or exact prompts used) because they claim it would be misused for evil. In other words, "trust me bro."
You are American, although you've discussed Ryanair before, which isn't exactly American. You have a number of comments and posts about Japan, which is strange, although you do drive a Japanese car.
Anyone who says that they can maintain perfect opsec over an extended period of time is seriously mistaken. A sufficiently motivated investigator with enough resources will join the dots eventually. The would-be evader has to be lucky every time whereas the investigator only has to be lucky once.
I'm pretty sure they can use the meta data the pull from your various interactions with search and the text you post online. These services build fingerprints of your habits using these techniques to follow you everywhere. At some point in the chain they could easily connect this fingerprint to your identity as soon as you log into and account that contains a piece of identifying information about you. The threat is real. I can foresee someone programming a terminal or app that obfuscates online behavior to avoid this fingerprinting in the future.
Unless I am misreading something. Take a look at surveillance capitalism to see what's possible right now. It's going to be 100x worse as LLMs become more advanced.
It's not the things you post online, it's the nuances behind the way you type and other ways to determine behavior that allows them to be able to build these kinds of profiles.
From what I can tell, the article/paper in question does not appear to utilize any of the techniques you mention, but I'd be interested to learn more about it.
> it's the nuances behind the way you type
I found this paper which talks about some of those methods.
The big companies that sell prediction products to advertisers. Google, Amazon, Microsoft, Apple, Meta... all of them are involved. I didn't read the paper but this is a known method they have been using for awhile to track people across sites of many types from social networking to online shopping sites
They refer to JP and language often enough in their search history and they state they are an american, and on 5G internet. I think working beyond this is doxxing. They could be anywhere.
I skimmed some of your comments, You seem to be in the US, at least mid30s, you bought a .dev domain and run your own email? I would think those are possible leads. You really don't think you slipped up once or twice in 5 years of posting? I think an llm would go through all your posts and context of the posts to get. and that would be easier to check if you used any other social media with the same name and see if the accounts have similarities.
I think Claude is guessing (educatedly - northern midwest does seem plausible). There's probably enough for the feds to track them down, but not me or an LLM.
I thought this would be more about stylometry but it's mostly about users literally posting the same identifiable information across multiple services, including in one example their age, dog name, profession.
It's all classic dox profiling techniques. Even the things like spelling differences being regional signals and commonality to specific things being discussed.
It's why one has to think about what is being posted to which community if using different identities, rather than posting the same things across all of them. Though any such effort would be a waste if reliant on some non-public info that later was exposed in a database breach which tied together previously unrelated profiles.
I’m curious if an LLM-based defense for this could be made. Like a browser plugin that warns you if you type identifiable information (like occupation) into a text field, and highlights turns of phrase that are “unusual” enough to be identifiable.
There was a tool shared here that could show which accounts belong to the same person based on the writing patterns. Can't remember the name, but it found my old accounts on HN pretty accurately.
Hnprofile.com which has since closed down - lettergram was the author - https://news.ycombinator.com/item?id=17942981
The internet is getting less interesting by the day.
The future is offline.
*selfhosted
*analog
*doomed
Anonymous account unmasking represents a new threat to anonymity. not just this technique with llms, but the earlier text similarity one.
But I think it would be generally easier to counter in the same way.
Use an llm or heuristics to pose as someone else.
not only do you erase your traces, you add false positives in to the system which reduces the overall effectiveness of these techniques in the future. A bit of poisoning the well.
I hope eventually an easy to use tool, with maybe a small local llm, can make it easy enough to do this, so that any future deanonymization attacks would be too untrustworthy to rely on
To state the obvious, we all need person, local tools to warn us when we’re making opsec errors.
There's a default Unix tool for that: https://www.man7.org/linux/man-pages/man1/yes.1.html
(Above 99% accuracy)
As a 32 year old Ghanaian woman living in Luang Prabang and studying as an ophthalmologist, this gives me some food for thought!
[dupe] Discussion on source: https://news.ycombinator.com/item?id=47139716
> If you request deletion of your Hacker News account, note that we reserve the right to refuse to (i) delete any of the submissions, favorites, or comments you posted on the Hacker News site
Probably not GDPR-compliant then if comments can be deanonymised by LLMs.
This is probably the worst piece of policy on whole HN. It has a evil feel to it. If HN wasn't so interesting/valueable, this would be the single reason NOT to use it at all.
All these comments live forever in HN datasets that people download anyway
My understanding is that the GDPR “right to be forgotten” applies to personal data. Are publicly available comments considered personal data?
Well, the username attached to them would surely be.
If they can help to deanonymize you, they must contain something personal. Writing pattern are pretty personal, certain spelling errors too, or the choose of words.
Absolutely anything relating to an anonymous person could help deanonymization, so that implies that anything relating to any person is personal data. Is that the GDPR’s position?
From ico.org.uk: “ It is important to note that opinions and inferences are also personal data, maybe special category data, if they directly or indirectly relate to that individual”
From gdpr-info.eu: “ Subjective information such as opinions, judgements or estimates can be personal data.”
So yes. HN is in violation of the GDPR. I had already filed a complaint about this policy at my local GDPR authority.
Only if said users happen to commit OPSEC failures themselves. LLMs aren't magic...
If someone can figure out who I am or what city I live in just by this username or my comments (with proof), I'll personally send you 500,000 JPY. I'm quite confident that's not going to happen though.
The paper referenced in the article does not even explain their exact testing methodology (such as the tools or exact prompts used) because they claim it would be misused for evil. In other words, "trust me bro."
Also see the previous discussion here: https://news.ycombinator.com/item?id=47139716
You are American, although you've discussed Ryanair before, which isn't exactly American. You have a number of comments and posts about Japan, which is strange, although you do drive a Japanese car.
A JDM car, probably, to be precise. I think they lived in Japan for at least a little while, e.g.: https://news.ycombinator.com/item?id=44679406#44686142
Anyone who says that they can maintain perfect opsec over an extended period of time is seriously mistaken. A sufficiently motivated investigator with enough resources will join the dots eventually. The would-be evader has to be lucky every time whereas the investigator only has to be lucky once.
You live on Earth. Now that I won let’s go double or nothing. I bet I can guess where you got dem shoes at.
He got them on his feet? He got them on the street?
I'm pretty sure they can use the meta data the pull from your various interactions with search and the text you post online. These services build fingerprints of your habits using these techniques to follow you everywhere. At some point in the chain they could easily connect this fingerprint to your identity as soon as you log into and account that contains a piece of identifying information about you. The threat is real. I can foresee someone programming a terminal or app that obfuscates online behavior to avoid this fingerprinting in the future.
Unless I am misreading something. Take a look at surveillance capitalism to see what's possible right now. It's going to be 100x worse as LLMs become more advanced.
It's not the things you post online, it's the nuances behind the way you type and other ways to determine behavior that allows them to be able to build these kinds of profiles.
Who is they? Which services?
From what I can tell, the article/paper in question does not appear to utilize any of the techniques you mention, but I'd be interested to learn more about it.
> it's the nuances behind the way you type
I found this paper which talks about some of those methods.
https://www.audiolabs-erlangen.de/content/04_fraunhofer/assi...
For example the "Text" section on page 91.
The big companies that sell prediction products to advertisers. Google, Amazon, Microsoft, Apple, Meta... all of them are involved. I didn't read the paper but this is a known method they have been using for awhile to track people across sites of many types from social networking to online shopping sites
With low precision, you're in Japan. But I don't need the JPY. of course that could be obfuscation.
The currency is not related to my location, I picked a random one, but thanks anyway :)
They said low precision. That might mean Spain, the US, etc
They refer to JP and language often enough in their search history and they state they are an american, and on 5G internet. I think working beyond this is doxxing. They could be anywhere.
Someone took the bait
What does 'of course that could be obfuscation' mean to you? because it doesn't mean 'took the bait' to me.
Ok boomer
Everyone commits opsec failures eventually. With LLMs linking anonymous accounts it just makes it even more likely to be caught.
You are ranger_danger
I skimmed some of your comments, You seem to be in the US, at least mid30s, you bought a .dev domain and run your own email? I would think those are possible leads. You really don't think you slipped up once or twice in 5 years of posting? I think an llm would go through all your posts and context of the posts to get. and that would be easier to check if you used any other social media with the same name and see if the accounts have similarities.
40 year old software dev in Detroit Michigan?
Not that I care, and that could be wildly off, but opsec is a wide term… and Claude one shot that… so safe out there bro, AI is wild
I think Claude is guessing (educatedly - northern midwest does seem plausible). There's probably enough for the feds to track them down, but not me or an LLM.
One solution is to flood the network with LLM slop and hide among the noise.
slop-steganography is that a name || a verb ?