6 ms·

“Do you use personal / behavioral data / usage to shape the user experience and recommendations, creating a reinforcing loop?” Roughly that.

by meowkit 2mo ago

matrix87 2mo ago
Okay but suppose you have an AI chatbot which takes in context from previous chats on what your interests are. So people just keep talking with the chatbot more and more as it gains more context. This would meet the criteria. It's not necessarily addictive though because you can stop talking with the chatbot at any time. Maybe it's something to do with the text-based interface vs short form video. Short form video is better at activating the lizard brain
- oblio 2mo ago
  > It's not necessarily addictive though because you can stop talking with the chatbot at any time. You can also stop taking drugs at any time. Or stop drinking. Or stop smoking. Or stop doomscrolling.
  - matrix87 2mo ago
    What I meant to say was, the standard for when we call something addictive is more hazy than that. In general, text based interfaces don't light up the brain in the same way. Psychologically it's a lot easier to walk away from a chatbot vs an infinite feed of short form video... potentially that difference could help find some kind of standard to determine addictiveness Edit: I wonder if a hypothetical digital experience like the samizdat out of infinite jest is possible? I.e. a digital experience which hits the brain in the perfect way to produce addiction without a chemical basis
    - ToucanLoucan 2mo ago
      > What I meant to say was, the standard for when we call something addictive is more hazy than that. If you as a person engage with a product to a degree where it causes demonstrable, measurable harm to yourself or your life, you are addicted.
      - matrix87 2mo ago
        If I cut my leg off with a chainsaw, am I addicted to the chainsaw? I think what matters is the intent of the design. E.g. if they design the product to make you need more of it (while it has an actively detrimental effect)
        ToucanLoucan 2mo ago
        > If I cut my leg off with a chainsaw, am I addicted to the chainsaw? If you repeatedly willingly do so despite the harm, then yes.
        blackqueeriroh 1mo ago
        You have moved the goalpost.
        UpsideDownRide 1mo ago
        Stop using buzzwords instead of engaging in a discussion. Someone clarifying their position is not necessairly moving goalposts, they might have just not laid out fully their idea in the first place.
        philipallstar 1mo ago
        But you still did it. If I sunbathe every day without sun cream, whom do we regulate? Assuming people have no ability to choose.
        stevesimmons 1mo ago
        For sunbathing in particular, do what Australia does: - large scale public health campaigns on dangers of skin cancer. With a catchy song and slogan ("slip, slop, slap... Slip on a shirt, slop on sunscreen, slap on a hat") - schools make it compulsory for kids to have hats in summer - regulate tanning salons, promote fake tans in preference to sunbeds - doctors check people for skin cancers during regular checkups And over just a few years, being very brown quickly gets associated with being reckless and unhealthy. In other words, regulate and educate
- phreack 2mo ago
  I'm not sure about that example. I'm seeing non tech people getting very addicted to their customized chat bots. Companies are very primed to eventually nudge them to enshittify the content of the responses and it could get bad.
- mslt 2mo ago
  You’ve not already noticed that the end of every consumer-facing LLM chat always includes an invitation for just one more turn to keep the loop going?
  - bad_haircut72 1mo ago
    No matter how intelligent they make the AIs they work for openAI/Anthropic, not the users. Our future lies in open, local models, or we might aswell give up our computers altogether
- brookst 2mo ago
  Law is about intent. It is not code, and it’s a mistake to think of law like code where you need to unit test and cover all edge cases. In law, it is sufficient to show an intent to take actions that foreseeably caused harm. Specific laws, of course but you don’t need to cover hypotheticals that capture all abstract cases in order to to show a particular actor caused harm.
- fwipsy 2mo ago
  Suppose you have a video service which takes context from previous videos you watched to show you more videos you're interested in. Doesn't sound very addictive, does it? You described it in an innocuous way, but I can absolutely see that being a lead-in to a parasocial "relationship" or even AI psychosis. > you can stop talking with the chatbot at any time. doubt. Lonely people will get addicted.
- solid_fuel 1mo ago
  > Okay but suppose you have an AI chatbot which takes in context from previous chats on what your interests are. So people just keep talking with the chatbot more and more as it gains more context. This would meet the criteria. It's not necessarily addictive though because you can stop talking with the chatbot at any time. You can stop scrolling instagram "at any time", too. Chatbots are actually a pretty poor but very relevant choice of example: they are already tuned with RLHF to be engaging and addictive to talk to. That's a huge reason why LLM psychosis is a thing, and it's a big reason why some people like using chatbots so much. They're designed to be addictive, just like facebook/instagram/tiktok are.
- rockskon 1mo ago
  "You can stop smoking at any time" "You can stop gambling at any time" As much as I may agree with you, you need a different argument.
  - Dylan16807 1mo ago
    I don't see why. The problem with your examples is that it's a person lying to themselves. GP is acting as an objective third party and telling the truth about their hypothetical. There are lots of things you really can stop at any time. I could stop eating peanut butter at any time, if I had a compelling reason.
    - rockskon 1mo ago
      Because if you want to convince someone away from believing it's an addiction so severe it needs an act of government to address, presenting an argument that all-but-panders to their "people spend so much time on it that it must be an addiction!" priors is counter productive.
charcircuit 2mo ago
What product does not listen to how users use the product to design the first or later version of a product?
- rf15 2mo ago
  Most products don't spy on you?
  - itake 1mo ago
    According to "Build things that last" author, 90% of the businesses evaluated are "customer first", meaning business listen to the customer and improve their produce based on that feedback.
    - solid_fuel 1mo ago
      voluntary feedback - things like "I wish the customer service was faster" - is not the same thing as measuring engagement time down to the millisecond and using that data to feed a presentation loop.
      - charcircuit 1mo ago
        Engagement is measuring voluntary feedback. If you aren't engaged you typically either move to the next piece of content or close the app. These are the voluntary mechanisms you as a customer report how good the recommendation is.
        solid_fuel 1mo ago
        No, engagement is not voluntary feedback, because there is no way to use the product without it being measured.
        charcircuit 1mo ago
        The user chooses themselves what part of a product they engage with. It's like reporting a post would be considered voluntary feedback on the anti spam systems even if it's the case that every report is measured and fed back into the model for predicting reports.
thesmtsolver2 2mo ago
So why leave out food industry which has provably caused health harm and killed people. Leaving out the modern food industry but going after social media shows that this is just politics as the food industry wets the beak and greases the plan generously on both sides of the aisle.
- dEnigma 2mo ago
  Unlike big tech, which would never do that.
- darthoctopus 2mo ago
  why do you think that implies leaving out the modern food industry, rather than suggesting that they, too, should be held accountable for flooding us with hyperpalatable foods to addict us? After all, they were literally engineered by the same people as Big Tobacco [1] [1]: https://www.washingtonpost.com/wellness/2023/09/19/addiction-foods-hyperpalatable-tobacco/ https://www.washingtonpost.com/wellness/2023/09/19/addiction...
- brookst 2mo ago
  In fully supportive of prosecuting wrongdoing in the food industry. But it seems strange to use belief in widespread wrongdoing in the food industry to oppose punishing wrongdoing in the tech industry. Is it just nihilism? All industries are corrupt, therefore prosecute none of them?
  - weatherlite 1mo ago
    It seems to me like Meta is being singled out here. It's likely not the start of a movement to change how we live and get rid of/diminish bad influences on our lives - because I don't think we wanna really do that - not for economic reasons and also values - we like our personal freedoms. This is simply targeting one company whose CEO is very disliked so we can all carry on with our lives and feel better about ourselves. I do support restricting social media (not just Meta) in general for people under 18.
    - 1mo ago
s1artibartfast 1mo ago
But isn't that just product design and Improvement? I think you have to figure out when this is in service of the user's intentions or not, and that's where it gets tricky. Apparently you can't trust users revealed preferences and have to use something else.
Aurornis 1mo ago
This would include everything from my Spotify recommending new music that I like to my Netflix showing me similar shows. I'm a responsible adult. I want those things. I make use of them. They make my life better when I'm trying to listen to some music while I work on something or when I have exactly 2 hours of TV watching time in a week and don't want to waste it finding a new show. I do not want regulations coming along and making all of my products worse due to heavy-handed regulation. Practically, it won't happen. These thought experiments are exercises where people imagine the websites they don't want other people using getting worse, while failing to imagine any way that it might negatively impact any site they use. Usually when I press the topic it turns into a game of whack-a-mole where the other people come up with more and more exceptions and specifics to these theoretical laws to reveal what they really have in mind, which is targeting some small handful of social media sites that they don't want other people to use. We see this with the age restriction laws where the comments come out in full force demanding ID checks and huge fines for Facebook and TikTok, but there are meltdowns whenever people realize that the regulations would be coming for their Discords, YouTube viewing, and Reddit access too. I halfway imagine this whole exercise ending with some big show about warnings, where users who join TikTok or other websites have to click a big warning sign that says they might be addictive. If some countries start instituting laws against recommendations or personalize recommendations, the popularity of VPNs will continue to rise so people can simply use the products like they want instead of having to deal with nerfed versions to satisfy some country's laws.
- dudefeliciano 1mo ago
  Ok then at the very least allow for a toggle to turn those algorithmic “recommendations” on and off
- Groxx 1mo ago
  Seems like you don't have to rule it out completely? "You might also like this show" is rather different than YouTube's "OMFG SHORTS WATCH THEM NOW" experience. Even if the line is fuzzy, I think many can agree some things are closer than others, and that's enough for some laws (there are plenty of other not-perfectly-testable laws out there as counterexamples).
mook 1mo ago
I think a better take is: > Do you use personal / behavioral data / usage to shape the user experience and recommendations? You do not get covered by Section 230. If they get to exercise editorial powers, then it doesn't make sense for them to be protected from liability for choosing what to publish. If they just give you a giant firehose, then they're not doing anything editorially and they should not have any liability as clearly that's not _their_ speech.
- lifeisstillgood 1mo ago
  This - one hundred times this. No matter what the “algorithm” is, at some point it’s a choice of what to put in front of the user - a publishing choice. Unless the rules are solely determined by myself and deterministic (no selection or choice on the platform end) then they are publishers. And liable.
- tzs 1mo ago
  That's not at all how Section 230 works. There's a popular myth that it distinguishes between "publishers" (who make editorial decisions about what content is shown and how it is shown) and "platforms" that just host it with minimal interference, but that is in fact just a myth. However, I don't really see how Section 230 is even relevant here. What Section 230 protects them from is for being liable for what is in the individual user posts posted to Facebook. It essentially just makes it so that if you have a beef with what someone posts on Facebook and want to sue you sue the poster rather than Facebook. What is being alleged is that Facebook's decisions in how to arrange and present and show or not show that user content has harmed people. Those are things that are not in scope for 230.
  - HappMacDonald 1mo ago
    Yet it remains relevant that "being liable for one's editorial decisions" is a good idea, and that differentiating between platforms which do and which do not exercise autonomy over how information gets prioritized and sorted may very well be relevant to this discussion.
camillomiller 1mo ago
Mix this with "...of User-Generated Multimedia Content", and you might get closer. This excludes music platforms, for example (although, Band Camp and Soundcloud might have some issues with that, admittedly).

“Do you use personal / behavioral data / usage to shape the user experience and recommendations, creating a reinforcing loop?” Roughly that.

by meowkit 2mo ago

“Do you use personal / behavioral data / usage to shape the user experience and recommendations, creating a reinforcing loop?” Roughly that.

matrix87 2mo ago

Okay but suppose you have an AI chatbot which takes in context from previous chats on what your interests are. So people just keep talking with the chatbot more and more as it gains more context. This would meet the criteria. It's not necessarily addictive though because you can stop talking with the chatbot at any time. Maybe it's something to do with the text-based interface vs short form video. Short form video is better at activating the lizard brain

oblio 2mo ago
> It's not necessarily addictive though because you can stop talking with the chatbot at any time. You can also stop taking drugs at any time. Or stop drinking. Or stop smoking. Or stop doomscrolling.
- matrix87 2mo ago
  What I meant to say was, the standard for when we call something addictive is more hazy than that. In general, text based interfaces don't light up the brain in the same way. Psychologically it's a lot easier to walk away from a chatbot vs an infinite feed of short form video... potentially that difference could help find some kind of standard to determine addictiveness Edit: I wonder if a hypothetical digital experience like the samizdat out of infinite jest is possible? I.e. a digital experience which hits the brain in the perfect way to produce addiction without a chemical basis
  - ToucanLoucan 2mo ago
    > What I meant to say was, the standard for when we call something addictive is more hazy than that. If you as a person engage with a product to a degree where it causes demonstrable, measurable harm to yourself or your life, you are addicted.
    - matrix87 2mo ago
      If I cut my leg off with a chainsaw, am I addicted to the chainsaw? I think what matters is the intent of the design. E.g. if they design the product to make you need more of it (while it has an actively detrimental effect)
      - ToucanLoucan 2mo ago
        > If I cut my leg off with a chainsaw, am I addicted to the chainsaw? If you repeatedly willingly do so despite the harm, then yes.
        blackqueeriroh 1mo ago
        You have moved the goalpost.
        UpsideDownRide 1mo ago
        Stop using buzzwords instead of engaging in a discussion. Someone clarifying their position is not necessairly moving goalposts, they might have just not laid out fully their idea in the first place.
        philipallstar 1mo ago
        But you still did it. If I sunbathe every day without sun cream, whom do we regulate? Assuming people have no ability to choose.
        stevesimmons 1mo ago
        For sunbathing in particular, do what Australia does: - large scale public health campaigns on dangers of skin cancer. With a catchy song and slogan ("slip, slop, slap... Slip on a shirt, slop on sunscreen, slap on a hat") - schools make it compulsory for kids to have hats in summer - regulate tanning salons, promote fake tans in preference to sunbeds - doctors check people for skin cancers during regular checkups And over just a few years, being very brown quickly gets associated with being reckless and unhealthy. In other words, regulate and educate
phreack 2mo ago
I'm not sure about that example. I'm seeing non tech people getting very addicted to their customized chat bots. Companies are very primed to eventually nudge them to enshittify the content of the responses and it could get bad.
mslt 2mo ago
You’ve not already noticed that the end of every consumer-facing LLM chat always includes an invitation for just one more turn to keep the loop going?
- bad_haircut72 1mo ago
  No matter how intelligent they make the AIs they work for openAI/Anthropic, not the users. Our future lies in open, local models, or we might aswell give up our computers altogether
brookst 2mo ago
Law is about intent. It is not code, and it’s a mistake to think of law like code where you need to unit test and cover all edge cases. In law, it is sufficient to show an intent to take actions that foreseeably caused harm. Specific laws, of course but you don’t need to cover hypotheticals that capture all abstract cases in order to to show a particular actor caused harm.
fwipsy 2mo ago
Suppose you have a video service which takes context from previous videos you watched to show you more videos you're interested in. Doesn't sound very addictive, does it? You described it in an innocuous way, but I can absolutely see that being a lead-in to a parasocial "relationship" or even AI psychosis. > you can stop talking with the chatbot at any time. doubt. Lonely people will get addicted.
solid_fuel 1mo ago
> Okay but suppose you have an AI chatbot which takes in context from previous chats on what your interests are. So people just keep talking with the chatbot more and more as it gains more context. This would meet the criteria. It's not necessarily addictive though because you can stop talking with the chatbot at any time. You can stop scrolling instagram "at any time", too. Chatbots are actually a pretty poor but very relevant choice of example: they are already tuned with RLHF to be engaging and addictive to talk to. That's a huge reason why LLM psychosis is a thing, and it's a big reason why some people like using chatbots so much. They're designed to be addictive, just like facebook/instagram/tiktok are.
rockskon 1mo ago
"You can stop smoking at any time" "You can stop gambling at any time" As much as I may agree with you, you need a different argument.
- Dylan16807 1mo ago
  I don't see why. The problem with your examples is that it's a person lying to themselves. GP is acting as an objective third party and telling the truth about their hypothetical. There are lots of things you really can stop at any time. I could stop eating peanut butter at any time, if I had a compelling reason.
  - rockskon 1mo ago
    Because if you want to convince someone away from believing it's an addiction so severe it needs an act of government to address, presenting an argument that all-but-panders to their "people spend so much time on it that it must be an addiction!" priors is counter productive.

charcircuit 2mo ago

What product does not listen to how users use the product to design the first or later version of a product?

rf15 2mo ago
Most products don't spy on you?
- itake 1mo ago
  According to "Build things that last" author, 90% of the businesses evaluated are "customer first", meaning business listen to the customer and improve their produce based on that feedback.
  - solid_fuel 1mo ago
    voluntary feedback - things like "I wish the customer service was faster" - is not the same thing as measuring engagement time down to the millisecond and using that data to feed a presentation loop.
    - charcircuit 1mo ago
      Engagement is measuring voluntary feedback. If you aren't engaged you typically either move to the next piece of content or close the app. These are the voluntary mechanisms you as a customer report how good the recommendation is.
      - solid_fuel 1mo ago
        No, engagement is not voluntary feedback, because there is no way to use the product without it being measured.
        charcircuit 1mo ago
        The user chooses themselves what part of a product they engage with. It's like reporting a post would be considered voluntary feedback on the anti spam systems even if it's the case that every report is measured and fed back into the model for predicting reports.

thesmtsolver2 2mo ago

So why leave out food industry which has provably caused health harm and killed people. Leaving out the modern food industry but going after social media shows that this is just politics as the food industry wets the beak and greases the plan generously on both sides of the aisle.

dEnigma 2mo ago
Unlike big tech, which would never do that.
darthoctopus 2mo ago
why do you think that implies leaving out the modern food industry, rather than suggesting that they, too, should be held accountable for flooding us with hyperpalatable foods to addict us? After all, they were literally engineered by the same people as Big Tobacco [1] [1]: https://www.washingtonpost.com/wellness/2023/09/19/addiction-foods-hyperpalatable-tobacco/ https://www.washingtonpost.com/wellness/2023/09/19/addiction...
brookst 2mo ago
In fully supportive of prosecuting wrongdoing in the food industry. But it seems strange to use belief in widespread wrongdoing in the food industry to oppose punishing wrongdoing in the tech industry. Is it just nihilism? All industries are corrupt, therefore prosecute none of them?
- weatherlite 1mo ago
  It seems to me like Meta is being singled out here. It's likely not the start of a movement to change how we live and get rid of/diminish bad influences on our lives - because I don't think we wanna really do that - not for economic reasons and also values - we like our personal freedoms. This is simply targeting one company whose CEO is very disliked so we can all carry on with our lives and feel better about ourselves. I do support restricting social media (not just Meta) in general for people under 18.
  - 1mo ago

s1artibartfast 1mo ago

But isn't that just product design and Improvement? I think you have to figure out when this is in service of the user's intentions or not, and that's where it gets tricky. Apparently you can't trust users revealed preferences and have to use something else.

Aurornis 1mo ago

This would include everything from my Spotify recommending new music that I like to my Netflix showing me similar shows. I'm a responsible adult. I want those things. I make use of them. They make my life better when I'm trying to listen to some music while I work on something or when I have exactly 2 hours of TV watching time in a week and don't want to waste it finding a new show. I do not want regulations coming along and making all of my products worse due to heavy-handed regulation. Practically, it won't happen. These thought experiments are exercises where people imagine the websites they don't want other people using getting worse, while failing to imagine any way that it might negatively impact any site they use. Usually when I press the topic it turns into a game of whack-a-mole where the other people come up with more and more exceptions and specifics to these theoretical laws to reveal what they really have in mind, which is targeting some small handful of social media sites that they don't want other people to use. We see this with the age restriction laws where the comments come out in full force demanding ID checks and huge fines for Facebook and TikTok, but there are meltdowns whenever people realize that the regulations would be coming for their Discords, YouTube viewing, and Reddit access too. I halfway imagine this whole exercise ending with some big show about warnings, where users who join TikTok or other websites have to click a big warning sign that says they might be addictive. If some countries start instituting laws against recommendations or personalize recommendations, the popularity of VPNs will continue to rise so people can simply use the products like they want instead of having to deal with nerfed versions to satisfy some country's laws.

dudefeliciano 1mo ago
Ok then at the very least allow for a toggle to turn those algorithmic “recommendations” on and off
Groxx 1mo ago
Seems like you don't have to rule it out completely? "You might also like this show" is rather different than YouTube's "OMFG SHORTS WATCH THEM NOW" experience. Even if the line is fuzzy, I think many can agree some things are closer than others, and that's enough for some laws (there are plenty of other not-perfectly-testable laws out there as counterexamples).

mook 1mo ago

I think a better take is: > Do you use personal / behavioral data / usage to shape the user experience and recommendations? You do not get covered by Section 230. If they get to exercise editorial powers, then it doesn't make sense for them to be protected from liability for choosing what to publish. If they just give you a giant firehose, then they're not doing anything editorially and they should not have any liability as clearly that's not _their_ speech.

lifeisstillgood 1mo ago
This - one hundred times this. No matter what the “algorithm” is, at some point it’s a choice of what to put in front of the user - a publishing choice. Unless the rules are solely determined by myself and deterministic (no selection or choice on the platform end) then they are publishers. And liable.
tzs 1mo ago
That's not at all how Section 230 works. There's a popular myth that it distinguishes between "publishers" (who make editorial decisions about what content is shown and how it is shown) and "platforms" that just host it with minimal interference, but that is in fact just a myth. However, I don't really see how Section 230 is even relevant here. What Section 230 protects them from is for being liable for what is in the individual user posts posted to Facebook. It essentially just makes it so that if you have a beef with what someone posts on Facebook and want to sue you sue the poster rather than Facebook. What is being alleged is that Facebook's decisions in how to arrange and present and show or not show that user content has harmed people. Those are things that are not in scope for 230.
- HappMacDonald 1mo ago
  Yet it remains relevant that "being liable for one's editorial decisions" is a good idea, and that differentiating between platforms which do and which do not exercise autonomy over how information gets prioritized and sorted may very well be relevant to this discussion.

camillomiller 1mo ago

Mix this with "...of User-Generated Multimedia Content", and you might get closer. This excludes music platforms, for example (although, Band Camp and Soundcloud might have some issues with that, admittedly).