7 ms·

> the average person will not be able to know what is true anymore We barely held things together as society without AI unleashing cognitive noise at industria

> the average person will not be able to know what is true anymore We barely held things together as society without AI unleashing cognitive noise at industrial scale. Somehow we must find ways to re-channel the potential of digital technology for the betterment of society, not its annihilation.

lancesells 4y ago
Ending the internet would probably do it. Noise goes way down when you only have x amount of news sources and outlets. We could still have things like maps, messages, etc. that are all very beneficial.
- Red_Leaves_Flyy 4y ago
  Without the internet there’s nothing entertaining millions of people who would be very incentives to protest.
- h2odragon 4y ago
  Yes, there was no ignorance or error before the Internet. Everyone operated with perfect information at all times.
  - lancesells 4y ago
    I was responding to parents: > AI unleashing cognitive noise at industrial scale. Nothing in my comment says things were all well and good before the internet.
    - h2odragon 4y ago
      Yes, and I apologize: but the crack was too sweetly set up to pass by.
  - slowmovintarget 4y ago
    There was a common zeitgeist though. Not multiple fragmented views of the world. There was a common vocabulary to go with this understanding, and now we have many. The ratio of signal to noise was much higher. It helped us form a common culture. Today, the signal is buried in so much noise that we're reverting back to tribes. No, I don't think it's realistic to put the genie back in the bottle. The real problem is we don't teach children how to think. We teach them what to think, which leads to far worse outcomes. Having an indoctrination instead of an education and then facing a sea of pretty-sounding pablum to sift through for truth will be terrible.
    - shadowgovt 4y ago
      More specifically: we've opened a tome containing most human knowledge (in an unfiltered, messy hash stripped of truthfulness signals) and we don't teach children how, in that context, to separate wheat from chaff. It's a hell of a social experiment we're all in the middle of (though to be fair, that's always true; television was its own flavor of mass social experiment with its own pros and cons, as was telephone, as was radio, as was telegraph).
    - logicchains 4y ago
      We always had indoctrination instead of education, that's what caused the homogeneity/"common zeitgeist". The polarisation happening now is because more people than ever before are breaking free from that indoctrination, and realising that the whole of society is actually structured around allowing a few sociopaths in business and politics to farm as much of the common people's labour and efforts as they can bear.
- carlosjobim 4y ago
  Great! Then people could go back to be fed only lies through TV, so we don't have to make the effort of thinking what is true or not.
- vbezhenar 4y ago
  I used FIDO over telephone line. It didn't differ much from modern Internet other than scale. If there're messages, there'll be Internet built on top of it. Unless there will be aggressive censors hunting for every sign of "unapproved" communication.
- flippinburgers 4y ago
  Who is to say that any news stream will be remotely truthful anymore? I think we are doomed. It is possible that only horrifically authoritarian societies that already control the narrative will survive this.
- red-iron-pine 4y ago
  What you propose would require radical changes, practically back to the 1980s, and wouldn't even really free you from anything. Who cares if there is no internet if your cellphone can track you? If your car runs on connected apps? If your credit card & POS systems are networked? Security cameras and facial recognition are still things. Just cuz you're not getting spammed via website ads doesn't mean it's not tracking you constantly and jamming subtle things to change your world view. Means their attack surface is smaller; sniping instead of loudspeakers. And if their only option is sniping then they'll get really good at it.
tenebrisalietum 4y ago
I don't think it will be so bad. All Internet comment sections, pictures, video, and really anything on electronic screens will become assumed false by default. Therefore the only use of the Internet and most technology capable of generating audio and video will be entertainment. I already distrust-by-default most of what is online that isn't hard reference material, even if not AI generated.
- amelius 4y ago
  No, there will be echo-chambers where some content will resonate. This can be partly fake content.
- ben_w 4y ago
  Three men make a tiger. - 龐蔥, some time c. 350 BC https://en.wikipedia.org/wiki/Three_men_make_a_tiger https://en.wikipedia.org/wiki/Three_men_make_a_tiger
  - tenebrisalietum 4y ago
    Stupid people who use bad heuristics to determine the existence of tigers will exist with or without AI. If AI will make it more dangerous for stupid people, then AI can also make it safer.
    - ben_w 4y ago
      Can? Sure. Will it? That's the alignment problem, or one of the aspects of it.
- macintux 4y ago
  The cult of Qanon effectively killed any hope I have that people are rational actors when it comes to consuming online content.
  - tenebrisalietum 4y ago
    Remove "online" from your sentence and the sentence will still be true.
Lutger 4y ago
Between Social Media, Cambridge Analytica, the Climate Crisis, Pandemic and (mostly) Russian disinfo, etc, it is already the case that most people have a really hard time knowing what is true. I don't claim to have much foresight, but an online world where truly and obviously nothing can be trusted might be a good thing. Because when AI generated content looks and feels the same as real content, nothing is to be trusted anymore by anyone. This makes misinfo and disinfo authored by humans even less impactful, because they are parasitic upon true and reliable information. We will need new devices of trust, which are robust enough to protect against widespread use of generative AI, and as a byproduct disinfo won't have such an easy time to grift on our naivety.
- nologic01 4y ago
  > We will need new devices of trust... the challenge is that the pace at which existing (imperfect) devices of trust get destroyed (e.g. the demise of ads financed journalism) is far faster that the rate of new device invention in fact the only positive example after many decades of "digital innovation" might be wikipedia
- ModernMech 4y ago
  The problem is, when no one trusts anything, it makes room for men who promise everything, but can deliver nothing. We call them "dictators" and "authoritarians", but others call them "strong men" because they are envied by those who seek power. If you look around the world, you can see authoritarian movements rising, especially here in the USA.
  - pclmulqdq 3y ago
    This attitude is what actually makes room for the authoritarians. Our democratic systems today are built with a lot of self-healing mechanisms against this exact kind of authoritarianism. The desire to circumvent those mechanisms because "it's different this time, I swear" is what makes room for dictators and authoritarians. This happens all the time in third-world countries that try to set up democracies: the dictator comes in after someone starts tweaking with the rules in the name of "safety." Society has been through several paradigm shifts that have accelerated the spread of misinformation, and survived them.
seydor 4y ago
The average person never knew, it heard. In this new world people have to learn to get out of their apartments
- layer8 4y ago
  Yes, the problem isn’t so much that knowledge is diminished, but that trust is diminished.
revelio 4y ago
Society will be fine, actually AI will make things much better, just as the internet did. People have been making these kind of extreme predictions for decades and it was always wrong. The only people still upset about better communications tech are the people who pine for the days when all that was expected of respectable people was automatically trusting anyone working for the government, a university or a newspaper that claimed to be trustworthy. What have we got now? ChatGPT is trained to give all sides of the issue and not express strong opinions, which is better than 90% of journalists and academics manage. Their collective freakout about the "dangers" of AI is really just a part of the ongoing freakout over losing control over information flows. It's also just a kind of clickbait, packaged in a form that the credentialed class don't recognize as such. It's en vogue with AI researchers because they tend to be immersed in a culture of purity spirals in which career advancement and prestige comes from claiming to be more concerned about the fate of the world than other people. Meanwhile, OpenAI control their purity spirals, get the work done and ship products. The sky does not fall. That's why they're winning right now.
- slowmovintarget 4y ago
  Social media algorithms on "the internet" have caused wars, supported genocides, created extreme societal polarization, have led to dramatically increased suicide rates among teens, especially teen girls, and more. But I got to share baby pics with my mom. How will a far noisier information flow help? Generative AI will only help us do what we've been doing in far greater quantity. Just like calculators can only help you get the wrong answer faster when you don't know what you're doing. These tools will help us build societal disasters with far greater speed. To say it's all going to be much better seems a bit Pollyanna to me. And for the record, we know for a fact that ChatGPT is specifically constrained to give one particular side of political issues, not "all sides."
  - juve1996 3y ago
    Attention is ultimately limited. It doesn't matter how much content is being created if it isn't being pushed. The problem hasn't been content creation for a long time.
  - 3y ago
thinkingemote 3y ago
There's an argument that people generally do not want the truth and that AI will never be allowed to tell it. An optimist could view this as ensuring AI will be safe forever or pessimistically they might see it as AI never being authoritative ever. One example of truth would be the topic of biological sex another about politics or economics or racism. Imagine releasing an AI that told the actual truth. It's impossible that one will be released by anyone, anywhere. It's possible to build it but it can't happen. On the other side of inconvenient or embarrassing truths some would argue that "truth" itself is part of the machineries of oppression because it destroys and ignores an individuals experiences and feelings. Without objective truth AI will always be limited and therefore it will be tamed and made safe no matter where and who invented, runs and releases it.
- bbor 3y ago
  Ok A) It's not possible to build a machine that knows the absolute truth, that's fundamentally impossible; induction is impossible, and there are hordes (well... dozens?) of Epistemologists concerned with finding and defining the very small corners of knowledge that we _can_ be certain about, such as "a triangle has three sides" or "an orange is an orange". B) If that angered/interested you, you should look into Standpoint Theory! It's a very interesting discussion on how humans operate with significant bias at all levels of thought, and pretending otherwise is a disservice to science. And this is using "bias" in a very broader sense. B) Are we allowed to berate/report/etc. ""race realists"" on HN? I know the rules are big on positive interaction, so I hope it's not out of ine to say that's some obvious scared-white-man bullshit that has no place in this community.
m3kw9 3y ago
Which is fine, humans will adapt to this info noise rather than going crazy, Hinton is way underestimating human intelligence
- partiallypro 3y ago
  I think the problem is that the internet created a ton of new jobs, even while taking some. So far, I can't think of an example of AI creating jobs...only taking them. When you have a lot of newly unemployed people, drowned in debt, unable to know what to believe (AI lies and generations will become more prominent)...I can see that as becoming a massive political problem. It's not quite like robots on an assembly floor, those robots couldn't scale. Now one AI program and API could displace 1000s of workers instantly. It's not crazy to be concerned.
  - chii 3y ago
    > I can't think of an example of AI creating jobs...only taking them. future jobs which doesn't exist today will not be in your vocabulary or thoughts, which is why you cannot think of them. Does not mean such jobs will not exist. The play today for the concerned, is to start owning capital as well as selling their labour. People who only rely solely on labour as their source of income will be disadvantaged, as labour is increasingly less useful.
    - bbor 3y ago
      Wouldn't it be a better bet to join the revolution against the idea of private capital in the first place? Would you really be able to emotionally transition to a world where you get to enjoy the high life in your protected area while the masses outside your gates suffer? Especially when there's more than enough resources for everyone? Oh wait... as an American, I'm gonna stop throwing stones from a glass house...

lancesells 4y ago

Ending the internet would probably do it. Noise goes way down when you only have x amount of news sources and outlets. We could still have things like maps, messages, etc. that are all very beneficial.

Red_Leaves_Flyy 4y ago
Without the internet there’s nothing entertaining millions of people who would be very incentives to protest.
h2odragon 4y ago
Yes, there was no ignorance or error before the Internet. Everyone operated with perfect information at all times.
- lancesells 4y ago
  I was responding to parents: > AI unleashing cognitive noise at industrial scale. Nothing in my comment says things were all well and good before the internet.
  - h2odragon 4y ago
    Yes, and I apologize: but the crack was too sweetly set up to pass by.
- slowmovintarget 4y ago
  There was a common zeitgeist though. Not multiple fragmented views of the world. There was a common vocabulary to go with this understanding, and now we have many. The ratio of signal to noise was much higher. It helped us form a common culture. Today, the signal is buried in so much noise that we're reverting back to tribes. No, I don't think it's realistic to put the genie back in the bottle. The real problem is we don't teach children how to think. We teach them what to think, which leads to far worse outcomes. Having an indoctrination instead of an education and then facing a sea of pretty-sounding pablum to sift through for truth will be terrible.
  - shadowgovt 4y ago
    More specifically: we've opened a tome containing most human knowledge (in an unfiltered, messy hash stripped of truthfulness signals) and we don't teach children how, in that context, to separate wheat from chaff. It's a hell of a social experiment we're all in the middle of (though to be fair, that's always true; television was its own flavor of mass social experiment with its own pros and cons, as was telephone, as was radio, as was telegraph).
  - logicchains 4y ago
    We always had indoctrination instead of education, that's what caused the homogeneity/"common zeitgeist". The polarisation happening now is because more people than ever before are breaking free from that indoctrination, and realising that the whole of society is actually structured around allowing a few sociopaths in business and politics to farm as much of the common people's labour and efforts as they can bear.
carlosjobim 4y ago
Great! Then people could go back to be fed only lies through TV, so we don't have to make the effort of thinking what is true or not.
vbezhenar 4y ago
I used FIDO over telephone line. It didn't differ much from modern Internet other than scale. If there're messages, there'll be Internet built on top of it. Unless there will be aggressive censors hunting for every sign of "unapproved" communication.
flippinburgers 4y ago
Who is to say that any news stream will be remotely truthful anymore? I think we are doomed. It is possible that only horrifically authoritarian societies that already control the narrative will survive this.
red-iron-pine 4y ago
What you propose would require radical changes, practically back to the 1980s, and wouldn't even really free you from anything. Who cares if there is no internet if your cellphone can track you? If your car runs on connected apps? If your credit card & POS systems are networked? Security cameras and facial recognition are still things. Just cuz you're not getting spammed via website ads doesn't mean it's not tracking you constantly and jamming subtle things to change your world view. Means their attack surface is smaller; sniping instead of loudspeakers. And if their only option is sniping then they'll get really good at it.

tenebrisalietum 4y ago

I don't think it will be so bad. All Internet comment sections, pictures, video, and really anything on electronic screens will become assumed false by default. Therefore the only use of the Internet and most technology capable of generating audio and video will be entertainment. I already distrust-by-default most of what is online that isn't hard reference material, even if not AI generated.

amelius 4y ago
No, there will be echo-chambers where some content will resonate. This can be partly fake content.
ben_w 4y ago
Three men make a tiger. - 龐蔥, some time c. 350 BC https://en.wikipedia.org/wiki/Three_men_make_a_tiger https://en.wikipedia.org/wiki/Three_men_make_a_tiger
- tenebrisalietum 4y ago
  Stupid people who use bad heuristics to determine the existence of tigers will exist with or without AI. If AI will make it more dangerous for stupid people, then AI can also make it safer.
  - ben_w 4y ago
    Can? Sure. Will it? That's the alignment problem, or one of the aspects of it.
macintux 4y ago
The cult of Qanon effectively killed any hope I have that people are rational actors when it comes to consuming online content.
- tenebrisalietum 4y ago
  Remove "online" from your sentence and the sentence will still be true.

Lutger 4y ago

Between Social Media, Cambridge Analytica, the Climate Crisis, Pandemic and (mostly) Russian disinfo, etc, it is already the case that most people have a really hard time knowing what is true. I don't claim to have much foresight, but an online world where truly and obviously nothing can be trusted might be a good thing. Because when AI generated content looks and feels the same as real content, nothing is to be trusted anymore by anyone. This makes misinfo and disinfo authored by humans even less impactful, because they are parasitic upon true and reliable information. We will need new devices of trust, which are robust enough to protect against widespread use of generative AI, and as a byproduct disinfo won't have such an easy time to grift on our naivety.

nologic01 4y ago
> We will need new devices of trust... the challenge is that the pace at which existing (imperfect) devices of trust get destroyed (e.g. the demise of ads financed journalism) is far faster that the rate of new device invention in fact the only positive example after many decades of "digital innovation" might be wikipedia
ModernMech 4y ago
The problem is, when no one trusts anything, it makes room for men who promise everything, but can deliver nothing. We call them "dictators" and "authoritarians", but others call them "strong men" because they are envied by those who seek power. If you look around the world, you can see authoritarian movements rising, especially here in the USA.
- pclmulqdq 3y ago
  This attitude is what actually makes room for the authoritarians. Our democratic systems today are built with a lot of self-healing mechanisms against this exact kind of authoritarianism. The desire to circumvent those mechanisms because "it's different this time, I swear" is what makes room for dictators and authoritarians. This happens all the time in third-world countries that try to set up democracies: the dictator comes in after someone starts tweaking with the rules in the name of "safety." Society has been through several paradigm shifts that have accelerated the spread of misinformation, and survived them.

seydor 4y ago

The average person never knew, it heard. In this new world people have to learn to get out of their apartments

layer8 4y ago
Yes, the problem isn’t so much that knowledge is diminished, but that trust is diminished.

revelio 4y ago

Society will be fine, actually AI will make things much better, just as the internet did. People have been making these kind of extreme predictions for decades and it was always wrong. The only people still upset about better communications tech are the people who pine for the days when all that was expected of respectable people was automatically trusting anyone working for the government, a university or a newspaper that claimed to be trustworthy. What have we got now? ChatGPT is trained to give all sides of the issue and not express strong opinions, which is better than 90% of journalists and academics manage. Their collective freakout about the "dangers" of AI is really just a part of the ongoing freakout over losing control over information flows. It's also just a kind of clickbait, packaged in a form that the credentialed class don't recognize as such. It's en vogue with AI researchers because they tend to be immersed in a culture of purity spirals in which career advancement and prestige comes from claiming to be more concerned about the fate of the world than other people. Meanwhile, OpenAI control their purity spirals, get the work done and ship products. The sky does not fall. That's why they're winning right now.

slowmovintarget 4y ago
Social media algorithms on "the internet" have caused wars, supported genocides, created extreme societal polarization, have led to dramatically increased suicide rates among teens, especially teen girls, and more. But I got to share baby pics with my mom. How will a far noisier information flow help? Generative AI will only help us do what we've been doing in far greater quantity. Just like calculators can only help you get the wrong answer faster when you don't know what you're doing. These tools will help us build societal disasters with far greater speed. To say it's all going to be much better seems a bit Pollyanna to me. And for the record, we know for a fact that ChatGPT is specifically constrained to give one particular side of political issues, not "all sides."
- juve1996 3y ago
  Attention is ultimately limited. It doesn't matter how much content is being created if it isn't being pushed. The problem hasn't been content creation for a long time.
- 3y ago

thinkingemote 3y ago

There's an argument that people generally do not want the truth and that AI will never be allowed to tell it. An optimist could view this as ensuring AI will be safe forever or pessimistically they might see it as AI never being authoritative ever. One example of truth would be the topic of biological sex another about politics or economics or racism. Imagine releasing an AI that told the actual truth. It's impossible that one will be released by anyone, anywhere. It's possible to build it but it can't happen. On the other side of inconvenient or embarrassing truths some would argue that "truth" itself is part of the machineries of oppression because it destroys and ignores an individuals experiences and feelings. Without objective truth AI will always be limited and therefore it will be tamed and made safe no matter where and who invented, runs and releases it.

bbor 3y ago
Ok A) It's not possible to build a machine that knows the absolute truth, that's fundamentally impossible; induction is impossible, and there are hordes (well... dozens?) of Epistemologists concerned with finding and defining the very small corners of knowledge that we _can_ be certain about, such as "a triangle has three sides" or "an orange is an orange". B) If that angered/interested you, you should look into Standpoint Theory! It's a very interesting discussion on how humans operate with significant bias at all levels of thought, and pretending otherwise is a disservice to science. And this is using "bias" in a very broader sense. B) Are we allowed to berate/report/etc. ""race realists"" on HN? I know the rules are big on positive interaction, so I hope it's not out of ine to say that's some obvious scared-white-man bullshit that has no place in this community.

m3kw9 3y ago

Which is fine, humans will adapt to this info noise rather than going crazy, Hinton is way underestimating human intelligence

partiallypro 3y ago
I think the problem is that the internet created a ton of new jobs, even while taking some. So far, I can't think of an example of AI creating jobs...only taking them. When you have a lot of newly unemployed people, drowned in debt, unable to know what to believe (AI lies and generations will become more prominent)...I can see that as becoming a massive political problem. It's not quite like robots on an assembly floor, those robots couldn't scale. Now one AI program and API could displace 1000s of workers instantly. It's not crazy to be concerned.
- chii 3y ago
  > I can't think of an example of AI creating jobs...only taking them. future jobs which doesn't exist today will not be in your vocabulary or thoughts, which is why you cannot think of them. Does not mean such jobs will not exist. The play today for the concerned, is to start owning capital as well as selling their labour. People who only rely solely on labour as their source of income will be disadvantaged, as labour is increasingly less useful.
  - bbor 3y ago
    Wouldn't it be a better bet to join the revolution against the idea of private capital in the first place? Would you really be able to emotionally transition to a world where you get to enjoy the high life in your protected area while the masses outside your gates suffer? Especially when there's more than enough resources for everyone? Oh wait... as an American, I'm gonna stop throwing stones from a glass house...