7 ms·

I don't think he is saying that encrypted communication should be illegal. He's saying that spreading lies through encrypted channels should be stopped. It woul

I don't think he is saying that encrypted communication should be illegal. He's saying that spreading lies through encrypted channels should be stopped. It would not be hard for WhatsApp to prevent people from spreading bogus videos, because they have access to the unencrypted information before and after sending. So either preventing people from sharing misinformation or preventing people from reading the misinformation. That's separate from the argument about whether or not it's a good thing (I wouldn't see a problem with it personally).

devcpp 6y ago
It would indeed be very hard, if WhatsApp's own claim that messages are encrypted end-to-end is true. Either way (and I disagree that this should be a separate discussion) all methods of global speech control are eventually used for evil.
- reillyse 6y ago
  So first off, I think "WhatsApp" is being used in two ways here. 1) is WhatsApp the company or the central server and the other 2) is the app on someone's phone. When people say end to end they mean from the WhatsApp app on someone's phone to the other WhatsApp app on another person's phone with no way WhatsApp the company can read the message. So what I'm saying is it's easy for the app on people's phone to filter or block sharing misinformation, because the app can see what the data is just like your eyes using the app can see what the data is. I'd also be fairly surprised if WhatsApp the app re-uploaded every single shared video, I'd say it's more likely they share a link or hash and encrypt that link in the message (but that is just a guess). To your second point. All methods of global speech are eventually used for evil. All methods of communication are eventually used for evil, all methods of food preparation are eventually used for evil. There are some flaws in that argument.
nojs 6y ago
> It would not be hard for WhatsApp to prevent people from spreading bogus videos The problem is there’s no agreed upon definition for what’s “bogus”. A while ago someone commented here with a list of statements ranging from obviously false obviously true to show how hard the problem is, I wish I could find it.
- hetspookjee 6y ago
  I wish you could too, as it sounds very interesting!
  - eanzenberg 6y ago
    Here’s some examples. A few weeks ago on HN was the outrage that Apple kept the 30% app store revenue cut on a return. This was outright false but led to a 1000 comment thread circle-jerk hating on Apple. Early in the pandemic the WHO claimed coronavirus could not be transmitted to humans. Having a central arbiter of “truth” is dangerous, fascist and there’s a reason freedom of speech has barely any limits on a constitutional basis.
Orou 6y ago
You'd think given the amount of time he's spent working on Coronavirus that he might see the distinction between treating the symptoms and treating the cause, so to speak. I've yet to see anyone provide any kind of evidence that suggests censorship changes people's minds about conspiracies - if anything it seems to be doing exactly the opposite.
- reillyse 6y ago
  I'm not sure there is evidence for or against. The argument I would make is not about censorship but preventing people from spreading lies. Even in the US where people hold free speech as some amazingly virtuous thing there are still curbs on speech, and what I'm saying is these limitations can be enforced in WhatsApp. Maybe you think that is censorship maybe you don't. Also on that point, the freedom of speech in the US is strongly correlated with conspiracy theories, if only because 1) you can freely spread malicious lies and 2) it can be in a lot of people's interests to do so. I think it's always been in people's interest to spread lies, but I think recently in the very partisan climate in the US the media (and I'm including social media there) have happened upon the discovery that spreading this misinformation is actually very very good for business. It generates outrage in the opposite camp and keeps one side of the divide entertained.
  - Orou 6y ago
    > The argument I would make is not about censorship but preventing people from spreading lies. Limiting speech is by definition censorship. There's no value judgement required as to the quality or veracity of that speech to determine whether or not restricting it counts as censorship. > the freedom of speech in the US is strongly correlated with conspiracy theories ...and water is highly correlated with drowning. We don't say 'water bad, less water'. Just because one thing is a prerequisite for something else doesn't mean it is responsible for that other thing. We have to look to, like you point out, the political climate, as well as the lack of trust in institutions, the education system, and many other factors. Taking a complex issue around how information is shared at an unprecedented speed and scale and saying "just ban it" is, I think, a shallow assessment.
_prototype_ 6y ago
And who defines what a "lie" is?
jlokier 6y ago
What an interesting idea! Several commenters point out end-to-end encryption would prevent filtering or tagging messages. But that's not true. The message analysis could be done at either endpoint without violating privacy. Tagging (or removing) a message before you send/forward it, or after you receive it, with "the central message of this comment has been tagged as "probably a hoax" by hoaxtracker.com; check out this CDC notice <here> to learn more". <here> does not need to be a URL which reveals much other than your general interest in the subject. But if that seems too revealing, it could already be already available as part of the endpoint's filtering data and readable locally. Lots of people forward (retweet), or write a little something before resharing what is false or misleading information, not realising they're doing so. I would not be surprised if getting those tags, rarely enough to stand out, before they send the message would cause some people to hesitate and check/think a bit more before sending. Maybe rephrase their attached comment into a question rather than confident outrage. Technically this is not much different from privacy-preserving spam filtering.
- imagiko 6y ago
  Whoa wait what..is it really E2E if they can analyse the content before putting it through their network?
  - jpsalm 6y ago
    Of course the party that is doing the encryption can see the original content...
  - panda88888 6y ago
    Yes, they can analyze it in the app on user devices before it is encrypted and transmitted. The app on user decide need access to clear text in order to encrypt the message. Same on the receiving end. The app on user device can analyze the message once it is decrypted on user device.
    - kshacker 6y ago
      Wow did I not read something similar last couple of days about on-device machine learning being better than in the cloud machine learning and how Apple got that right. That's where this may also be going then if we follow your train of thought.
1vuio0pswjnm7 6y ago
"It would be hard for WhatsApp to prevent people from spreading bogus videos, because they have access to the unencrypted information before and after sending." Facebook/WhatsApp claims WhatsApp messages are "end-to-end" encypted. For example, here https://webcache.googleusercontent.com/search?q=cache:https://www.whatsapp.com/security/ https://webcache.googleusercontent.com/search?q=cache:https:... Are the "end-to-end" encryption claims irrelevant if "they have access to the unecncypted information before and after sending"?
- usrusr 6y ago
  They could put a tiny truth ministry module in the app and remain e2e: "I'll be on my way home in five minutes!" [send] "Sorry, our algorithms have detected that you were about to spread misinformation. Please contact support to reactivate your account. Premium support is available to subscribers of our membership programme" I'm actually only half joking: while some local library code surely cannot tell truth from lies, that problem remains just as unsolved for arbitrary amounts of central effort. There's a reason truth ministries are a Bad Idea.
  - Erlich_Bachman 6y ago
    > I'm actually only half joking: while some local library code surely cannot tell truth from lies Hey, the way GPT(-2, -3, ...)s are going, maybe soon we can have an offline classifier of truth. /s.
amiga_500 6y ago
Surely American culture is the problem. Conspiracy nuts predate the internet. This is a deeper problem.
sukilot 6y ago
Then why did he mention encryption? > (I wouldn't see a problem with it personally The POTUS said CNN is fake news, hundreds of times. So, as you say, people should be prevented from sharing and reading it?! There's no problem with that?! The very first item on the USA Bill of Rights isn't important?
eanzenberg 6y ago
“Spreading lies.” Sure, when we have almighty God say what’s the truth and what isn’t, let us know.
sxxahsuxh 6y ago
You know, Snopes has said that running a fact checking organisation is incredibly difficult. It's easy to see why, you have to: 1) decide what the definition of truth is 2) differentiate between subjectivity and objectivity 3) differentiate between misleading and outright incorrect 4) investigate every piece of media thoroughly 5) avoid bias 6) peer review 7) correct any mistakes And this list is just off the top of my head. So what bill gates is saying is that you have to do all this at a scale of 1 billion users, all controlled roughly by a few centralised organisations. I think Bill's words are still kind of stupid even with context. The way to fight all of these problems has always been education, and social welfare, and things the government should ACTUALLY be doing, not vetting encryption schemes.
- reillyse 6y ago
  No one is saying we need to replace Snopes or emulate what they are doing. But if a platform like WhatsApp, Facebook or Twitter fact checked the most spread thousands videos per day, this problem would evaporate. You could do that with a team of 5 people. But you would need the will to do it, which is the part I think is lacking.
Canada 6y ago
The main purpose of E2EE communication between willing participants is that the content of the communication is not checked, inspected, scanned, questioned, sampled, matched, filtered, modified, blocked, altered, or otherwise interfered with in any way whatsoever except that which is explicitly configured and consented to by a party to the conversation. (eg. anti-virus, anti-spam, group membership, etc) That means no control of communication by the endpoint software either directly or indirectly between willing senders and recipients regardless of justification, and especially on the basis of whether or not data send from a willing sender to a willing receipient represents the "truth". The endpoint software vendor has no standing to judge that unless it's an opt-in anti-spam type feature.
bbarnett 6y ago
I'm not sure he even said that. When I look at the medium article, I see this: And when you have [posts] encrypted, there is no way to know what it is. I personally believe government should not allow those types of lies or fraud or child pornography [to be hidden with encryption like WhatsApp or Facebook Messenger]. Note the []? Was this added in by Medium? Gates as an edit after a read-and-OK-to-release? Did Gates have editorial input? Regardless, I believe [] means 'edited afterwards for clarity'. But by whom? I Googled, and found this: https://wccftech.com/bill-gates-hate-encryption/ https://wccftech.com/bill-gates-hate-encryption/ Yet in this case, there was no hatred of encryption, but picked quotes where he suggested that in one case, if you have the means, the government should be aided .. eg, a murderer. There was another article on hackernews, where many lamented how much the media just spins, takes quotes out of context, basically does whatever it wants to. I wonder, how much of this are we seeing here? (Note, Gates could very much be for back-doored encryption, but my point is, I don't think it's a clear position due to this medium article.. where that stances was in [] and added by someone after...)

devcpp 6y ago

It would indeed be very hard, if WhatsApp's own claim that messages are encrypted end-to-end is true. Either way (and I disagree that this should be a separate discussion) all methods of global speech control are eventually used for evil.

reillyse 6y ago
So first off, I think "WhatsApp" is being used in two ways here. 1) is WhatsApp the company or the central server and the other 2) is the app on someone's phone. When people say end to end they mean from the WhatsApp app on someone's phone to the other WhatsApp app on another person's phone with no way WhatsApp the company can read the message. So what I'm saying is it's easy for the app on people's phone to filter or block sharing misinformation, because the app can see what the data is just like your eyes using the app can see what the data is. I'd also be fairly surprised if WhatsApp the app re-uploaded every single shared video, I'd say it's more likely they share a link or hash and encrypt that link in the message (but that is just a guess). To your second point. All methods of global speech are eventually used for evil. All methods of communication are eventually used for evil, all methods of food preparation are eventually used for evil. There are some flaws in that argument.

nojs 6y ago

> It would not be hard for WhatsApp to prevent people from spreading bogus videos The problem is there’s no agreed upon definition for what’s “bogus”. A while ago someone commented here with a list of statements ranging from obviously false obviously true to show how hard the problem is, I wish I could find it.

hetspookjee 6y ago
I wish you could too, as it sounds very interesting!
- eanzenberg 6y ago
  Here’s some examples. A few weeks ago on HN was the outrage that Apple kept the 30% app store revenue cut on a return. This was outright false but led to a 1000 comment thread circle-jerk hating on Apple. Early in the pandemic the WHO claimed coronavirus could not be transmitted to humans. Having a central arbiter of “truth” is dangerous, fascist and there’s a reason freedom of speech has barely any limits on a constitutional basis.

Orou 6y ago

You'd think given the amount of time he's spent working on Coronavirus that he might see the distinction between treating the symptoms and treating the cause, so to speak. I've yet to see anyone provide any kind of evidence that suggests censorship changes people's minds about conspiracies - if anything it seems to be doing exactly the opposite.

reillyse 6y ago
I'm not sure there is evidence for or against. The argument I would make is not about censorship but preventing people from spreading lies. Even in the US where people hold free speech as some amazingly virtuous thing there are still curbs on speech, and what I'm saying is these limitations can be enforced in WhatsApp. Maybe you think that is censorship maybe you don't. Also on that point, the freedom of speech in the US is strongly correlated with conspiracy theories, if only because 1) you can freely spread malicious lies and 2) it can be in a lot of people's interests to do so. I think it's always been in people's interest to spread lies, but I think recently in the very partisan climate in the US the media (and I'm including social media there) have happened upon the discovery that spreading this misinformation is actually very very good for business. It generates outrage in the opposite camp and keeps one side of the divide entertained.
- Orou 6y ago
  > The argument I would make is not about censorship but preventing people from spreading lies. Limiting speech is by definition censorship. There's no value judgement required as to the quality or veracity of that speech to determine whether or not restricting it counts as censorship. > the freedom of speech in the US is strongly correlated with conspiracy theories ...and water is highly correlated with drowning. We don't say 'water bad, less water'. Just because one thing is a prerequisite for something else doesn't mean it is responsible for that other thing. We have to look to, like you point out, the political climate, as well as the lack of trust in institutions, the education system, and many other factors. Taking a complex issue around how information is shared at an unprecedented speed and scale and saying "just ban it" is, I think, a shallow assessment.

_prototype_ 6y ago

And who defines what a "lie" is?

jlokier 6y ago

What an interesting idea! Several commenters point out end-to-end encryption would prevent filtering or tagging messages. But that's not true. The message analysis could be done at either endpoint without violating privacy. Tagging (or removing) a message before you send/forward it, or after you receive it, with "the central message of this comment has been tagged as "probably a hoax" by hoaxtracker.com; check out this CDC notice <here> to learn more". <here> does not need to be a URL which reveals much other than your general interest in the subject. But if that seems too revealing, it could already be already available as part of the endpoint's filtering data and readable locally. Lots of people forward (retweet), or write a little something before resharing what is false or misleading information, not realising they're doing so. I would not be surprised if getting those tags, rarely enough to stand out, before they send the message would cause some people to hesitate and check/think a bit more before sending. Maybe rephrase their attached comment into a question rather than confident outrage. Technically this is not much different from privacy-preserving spam filtering.

imagiko 6y ago
Whoa wait what..is it really E2E if they can analyse the content before putting it through their network?
- jpsalm 6y ago
  Of course the party that is doing the encryption can see the original content...
- panda88888 6y ago
  Yes, they can analyze it in the app on user devices before it is encrypted and transmitted. The app on user decide need access to clear text in order to encrypt the message. Same on the receiving end. The app on user device can analyze the message once it is decrypted on user device.
  - kshacker 6y ago
    Wow did I not read something similar last couple of days about on-device machine learning being better than in the cloud machine learning and how Apple got that right. That's where this may also be going then if we follow your train of thought.

1vuio0pswjnm7 6y ago

"It would be hard for WhatsApp to prevent people from spreading bogus videos, because they have access to the unencrypted information before and after sending." Facebook/WhatsApp claims WhatsApp messages are "end-to-end" encypted. For example, here https://webcache.googleusercontent.com/search?q=cache:https://www.whatsapp.com/security/ https://webcache.googleusercontent.com/search?q=cache:https:... Are the "end-to-end" encryption claims irrelevant if "they have access to the unecncypted information before and after sending"?

usrusr 6y ago
They could put a tiny truth ministry module in the app and remain e2e: "I'll be on my way home in five minutes!" [send] "Sorry, our algorithms have detected that you were about to spread misinformation. Please contact support to reactivate your account. Premium support is available to subscribers of our membership programme" I'm actually only half joking: while some local library code surely cannot tell truth from lies, that problem remains just as unsolved for arbitrary amounts of central effort. There's a reason truth ministries are a Bad Idea.
- Erlich_Bachman 6y ago
  > I'm actually only half joking: while some local library code surely cannot tell truth from lies Hey, the way GPT(-2, -3, ...)s are going, maybe soon we can have an offline classifier of truth. /s.

amiga_500 6y ago

Surely American culture is the problem. Conspiracy nuts predate the internet. This is a deeper problem.

sukilot 6y ago

Then why did he mention encryption? > (I wouldn't see a problem with it personally The POTUS said CNN is fake news, hundreds of times. So, as you say, people should be prevented from sharing and reading it?! There's no problem with that?! The very first item on the USA Bill of Rights isn't important?

eanzenberg 6y ago

“Spreading lies.” Sure, when we have almighty God say what’s the truth and what isn’t, let us know.

sxxahsuxh 6y ago

You know, Snopes has said that running a fact checking organisation is incredibly difficult. It's easy to see why, you have to: 1) decide what the definition of truth is 2) differentiate between subjectivity and objectivity 3) differentiate between misleading and outright incorrect 4) investigate every piece of media thoroughly 5) avoid bias 6) peer review 7) correct any mistakes And this list is just off the top of my head. So what bill gates is saying is that you have to do all this at a scale of 1 billion users, all controlled roughly by a few centralised organisations. I think Bill's words are still kind of stupid even with context. The way to fight all of these problems has always been education, and social welfare, and things the government should ACTUALLY be doing, not vetting encryption schemes.

reillyse 6y ago
No one is saying we need to replace Snopes or emulate what they are doing. But if a platform like WhatsApp, Facebook or Twitter fact checked the most spread thousands videos per day, this problem would evaporate. You could do that with a team of 5 people. But you would need the will to do it, which is the part I think is lacking.

Canada 6y ago

The main purpose of E2EE communication between willing participants is that the content of the communication is not checked, inspected, scanned, questioned, sampled, matched, filtered, modified, blocked, altered, or otherwise interfered with in any way whatsoever except that which is explicitly configured and consented to by a party to the conversation. (eg. anti-virus, anti-spam, group membership, etc) That means no control of communication by the endpoint software either directly or indirectly between willing senders and recipients regardless of justification, and especially on the basis of whether or not data send from a willing sender to a willing receipient represents the "truth". The endpoint software vendor has no standing to judge that unless it's an opt-in anti-spam type feature.

bbarnett 6y ago

I'm not sure he even said that. When I look at the medium article, I see this: And when you have [posts] encrypted, there is no way to know what it is. I personally believe government should not allow those types of lies or fraud or child pornography [to be hidden with encryption like WhatsApp or Facebook Messenger]. Note the []? Was this added in by Medium? Gates as an edit after a read-and-OK-to-release? Did Gates have editorial input? Regardless, I believe [] means 'edited afterwards for clarity'. But by whom? I Googled, and found this: https://wccftech.com/bill-gates-hate-encryption/ https://wccftech.com/bill-gates-hate-encryption/ Yet in this case, there was no hatred of encryption, but picked quotes where he suggested that in one case, if you have the means, the government should be aided .. eg, a murderer. There was another article on hackernews, where many lamented how much the media just spins, takes quotes out of context, basically does whatever it wants to. I wonder, how much of this are we seeing here? (Note, Gates could very much be for back-doored encryption, but my point is, I don't think it's a clear position due to this medium article.. where that stances was in [] and added by someone after...)