5 ms·

After FB bought WhatsApp I’ve had multiple occasions where shortly after having conversations with people about products I haven’t searched for. The only place

After FB bought WhatsApp I’ve had multiple occasions where shortly after having conversations with people about products I haven’t searched for. The only place I’ve discussed it. Is in a whats app conversation. I got advertising in Facebook for those products. For example when looking for an apartment I told my agent (non business account) that I wanted safety catches on the windows. I immediately got advertising for safety catches and window gates. I don’t own cats but in a conversation with a friend who owns cats I said she should get one of those cat tree things and scratch poles. Right away Facebook starts showing me adverts for cat toys. This is not stuff I’ve searched or googled or anything. Just mentioned in WhatsApp. Maybe WhatsApp whats differently in America but in Singapore I get advertising in Facebook from conversations.

scrollaway 4y ago
None of this is evidence of decryption. At best it’s circumstantial anecdotes pointing to Facebook extracting topic metadata out of your convos. This can be done clientside, no decryption necessary. (And it’s easy to prove that it’s being done if that’s the case)
- klyrs 4y ago
  E2EE plus arbitrarily invasive clent-side sniffing/reporting should not be considered "E2EE" by any means. Even if reports to the mothership are strongly encrypted (at which point it becomes harder to determine what's in the reports).
  - scrollaway 4y ago
    I agree with you, but the claim that Facebook is decrypting and reencrypting things is a very very very severe one.
    - klyrs 4y ago
      I didn't claim they are. But they can. And without transparency, I have very little inclination to trust. I find it plausible that they'd scan for keywords and phone home when they get hits -- advertising businesses are not known for their scruples and fb has not distinguished itself in that regard. Once that machinery is in place, it's ripe for abuse. I don't claim any proof it's happening, but it would be naive to blindly trust that it isn't.
gruez 4y ago
That's... not really good evidence. Your story sounds almost identical to the "facebook/google is eavesdropping on me" stories that frequently make the rounds on popular discourse. Unfortunately, that's basically the Sasquatch of the privacy world (ie. there are many people with anecdotes claiming it exists, but very little in the way of actual evidence like network captures or decompiled binaries). If anything, it's worse than Sasquatch because at least with Sasquatch you could claim that there aren't many of them and/or they're actively avoiding humans so they're hard to photograph, but the "facebook/google is eavesdropping on me" stories implies it's happening to everyone so capturing an instance should only be a matter of technical skill rather than luck.
- philliphaydon 4y ago
  Can you explain why advertising shows only in facebook, for keywords only used inside an apparently private conversation? Unless WhatsApp is audited completely, not just looking at some source code but also how all information passes through facebook and back to a receiver. Then we can only assume the e2ee is just marketing fluff and not something that is done in favour of privacy.
  - gruez 4y ago
    > Can you explain why advertising shows only in facebook, for keywords only used inside an apparently private conversation? obvious explanations: 1. coincidence/luck + confirmation bias. this can also be aided by demographic/interest factors. For instance, suppose there are 10,000 "interests" that facebook tracks. You'd think that for a given conversation that contained such a topic, the chance that you'd see ads for it was 1 in 10,000. However, if the conversation topic was "mechanical keyboards", and facebook knew that you were male, 18-35, and are interested in video games, then it might be able to infer that you're much more likely to be interested in mechanical keyboards and therefore show you ads for mechanical keyboards 1 in 50 times rather than 1 in 10,000 times. 2. the other side ended up leaking the information to facebook (eg. through a search or clicking on an ad)
    - philliphaydon 4y ago
kaba0 4y ago
The person you talked about may have searched for the term, and knowing that you are friends facebook can easily assume that you would be also interested in this thing. It may also operate based on metadata of you have talked recently.

scrollaway 4y ago

None of this is evidence of decryption. At best it’s circumstantial anecdotes pointing to Facebook extracting topic metadata out of your convos. This can be done clientside, no decryption necessary. (And it’s easy to prove that it’s being done if that’s the case)

klyrs 4y ago
E2EE plus arbitrarily invasive clent-side sniffing/reporting should not be considered "E2EE" by any means. Even if reports to the mothership are strongly encrypted (at which point it becomes harder to determine what's in the reports).
- scrollaway 4y ago
  I agree with you, but the claim that Facebook is decrypting and reencrypting things is a very very very severe one.
  - klyrs 4y ago
    I didn't claim they are. But they can. And without transparency, I have very little inclination to trust. I find it plausible that they'd scan for keywords and phone home when they get hits -- advertising businesses are not known for their scruples and fb has not distinguished itself in that regard. Once that machinery is in place, it's ripe for abuse. I don't claim any proof it's happening, but it would be naive to blindly trust that it isn't.

gruez 4y ago

That's... not really good evidence. Your story sounds almost identical to the "facebook/google is eavesdropping on me" stories that frequently make the rounds on popular discourse. Unfortunately, that's basically the Sasquatch of the privacy world (ie. there are many people with anecdotes claiming it exists, but very little in the way of actual evidence like network captures or decompiled binaries). If anything, it's worse than Sasquatch because at least with Sasquatch you could claim that there aren't many of them and/or they're actively avoiding humans so they're hard to photograph, but the "facebook/google is eavesdropping on me" stories implies it's happening to everyone so capturing an instance should only be a matter of technical skill rather than luck.

philliphaydon 4y ago
Can you explain why advertising shows only in facebook, for keywords only used inside an apparently private conversation? Unless WhatsApp is audited completely, not just looking at some source code but also how all information passes through facebook and back to a receiver. Then we can only assume the e2ee is just marketing fluff and not something that is done in favour of privacy.
- gruez 4y ago
  > Can you explain why advertising shows only in facebook, for keywords only used inside an apparently private conversation? obvious explanations: 1. coincidence/luck + confirmation bias. this can also be aided by demographic/interest factors. For instance, suppose there are 10,000 "interests" that facebook tracks. You'd think that for a given conversation that contained such a topic, the chance that you'd see ads for it was 1 in 10,000. However, if the conversation topic was "mechanical keyboards", and facebook knew that you were male, 18-35, and are interested in video games, then it might be able to infer that you're much more likely to be interested in mechanical keyboards and therefore show you ads for mechanical keyboards 1 in 50 times rather than 1 in 10,000 times. 2. the other side ended up leaking the information to facebook (eg. through a search or clicking on an ad)
  - philliphaydon 4y ago

kaba0 4y ago

The person you talked about may have searched for the term, and knowing that you are friends facebook can easily assume that you would be also interested in this thing. It may also operate based on metadata of you have talked recently.