Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
×
The Internet AI Robotics Idle

The 'Dead Internet' Theory Posits Forums are Now Almost Entirely Overrun By AI (theatlantic.com) 147

Ideas from 4chan (including its paranormal section) have percolated into the "dead internet" theory, writes the Atlantic, with a seminal post on another forum by "IlluminatiPirate" now arguing that the internet is almost entirely overrun by artificial intelligence: Like lots of other online conspiracy theories, the audience for this one is growing because of discussion led by a mix of true believers, sarcastic trolls, and idly curious lovers of chitchat... Peppered with casually offensive language, the post suggests that the internet died in 2016 or early 2017, and that now it is "empty and devoid of people," as well as "entirely sterile." Much of the "supposedly human-produced content" you see online was actually created using AI, IlluminatiPirate claims, and was propagated by bots, possibly aided by a group of "influencers" on the payroll of various corporations that are in cahoots with the government. The conspiring group's intention is, of course, to control our thoughts and get us to purchase stuff... He argues that all modern entertainment is generated and recommended by an algorithm; gestures at the existence of deepfakes, which suggest that anything at all may be an illusion; and links to a New York story from 2018 titled "How Much of the Internet Is Fake? Turns Out, a Lot of It, Actually."

"I think it's entirely obvious what I'm subtly suggesting here given this setup," the post continues. "The U.S. government is engaging in an artificial intelligence powered gaslighting of the entire world population." So far, the original post has been viewed more than 73,000 times...

The theory has become fodder for dramatic YouTube explainers, including one that summarizes the original post in Spanish and has been viewed nearly 260,000 times. Speculation about the theory's validity has started appearing in the widely read Hacker News forum and among fans of the massively popular YouTube channel Linus Tech Tips. In a Reddit forum about the paranormal, the theory is discussed as a possible explanation for why threads about UFOs seem to be "hijacked" by bots so often. The theory's spread hasn't been entirely organic. IlluminatiPirate has posted a link to his manifesto in several Reddit forums that discuss conspiracy theories... Anyway ... dead-internet theory is pretty far out-there. But unlike the internet's many other conspiracy theorists, who are boring or really gullible or motivated by odd politics, the dead-internet people kind of have a point... [Y]ou could even say that the point of the theory is so obvious, it's cliché — people talk about longing for the days of weird web design and personal sites and listservs all the time. Even Facebook employees say they miss the "old" internet. The big platforms do encourage their users to make the same conversations and arcs of feeling and cycles of outrage happen over and over, so much so that people may find themselves acting like bots, responding on impulse in predictable ways to things that were created, in all likelihood, to elicit that very response.

That 2018 article in New York magazine had argued that (at that time) a majority of web traffic was probably coming from bots — including especially high bot traffic on YouTube — while even the engagement metrics for major sites like Facebook had been gamed or inflated.

But whether or not that's changed, the Atlantic shares a compelling argument from a forum poster arguing that their very presence in this discussion proves they must be a bot. "If I was real I'm pretty sure I'd be out there living each day to the fullest and experiencing everything I possibly could with every given moment of the relatively infinitesimal amount of time I'll exist for instead of posting on the internet about nonsense."
This discussion has been archived. No new comments can be posted.

The 'Dead Internet' Theory Posits Forums are Now Almost Entirely Overrun By AI

Comments Filter:
  • by hey! ( 33014 ) on Sunday September 05, 2021 @01:38PM (#61766157) Homepage Journal

    when natural stupidity is so abundant?

    • by Immerman ( 2627577 ) on Sunday September 05, 2021 @01:52PM (#61766211)

      Yeah, I don't believe it. If discussion forums really had been taken over by mindless bots, the general level of conversation would be much higher.

      • I'm not an AI expert, but from what I've concluded after coming across stories from time to time about AI generated works, is that AI starts to fuck up more and more the longer a generated work is, to the point of devolving into absolute nonsense.

        It may be able to generate one or two paragraphs which would fool most people, but after that it becomes real obvious that it was generated by a machine and not a human.

        • by sg_oneill ( 159032 ) on Sunday September 05, 2021 @08:53PM (#61767345)

          Basically yeah. Its the lack of a higher level goal to structure the larger tech that has it undone.

          GPT-3 is astonishing. It really can generate some amazing results, and it really does feel like theres an underlying "intelligence" driving it. I've seen the damn thing crack jokes whilst screwing about with it on AI Dungeon, and not jokes it would have scraped off the net either but highly contextual (I summoned the beatles into a dungeon, and "john lennon" asked "why would you ask the beatles to defend you when you could ask for a knight?" to which Ringo replied "If I of all people can be a great drummer, You can be a great Knight, John.".) . It really is impressive.

          But it became obvious after a while the thing has no continuity. It has a rough implementation of "whats going on right now" but theres no arc to it. Its more reminicient of the way children tell stories where its just a series of cool stuff.

          And that itself is impressive, an AI that thinks like a child (or at least does an excellent iimprssion of it) IS an achievement. But far from useful.

          • by Malays2 bowman ( 6656916 ) on Monday September 06, 2021 @12:00AM (#61767703)

            It's not too hard to make an AI system generate multiple short passages that all follow a common theme. But all attempts to make AI generate a long singular piece of work, such as a novel so far has failed.

              I doubt it would be very hard to make an AI system generate a long form novel that from beginning to end that makes sense, but the end result would be a dry, dull work because it can't come up with exciting and unexpected (yet coherent) plot twists on it's own, Anything like that that the program adds would be something that a human created and seeded the novel generating program with.

            • All AI really is is RND + decision tree + logging of previous decisions so it won't try to generate the same thing more than once and sometimes to add items to the decision tree (i.e. Do X because it's correct but don't do Y, etc...)

              • "All AI really is is RND + decision tree + logging of previous decisions "

                No. Thats how the old "Eliza" bots used to work. Its a symbolist approach to AI, which largely has been abandoned since the 1990s outside of Prolog style inference engines.

                GPT3 , which powers AI Dungeon, is a convolutional neural network that got fed basically the internet and then OpenAI stewed it in a few million dollars worth of GPU Tensorflow time on AWS.

                Theres no RNG, Decision tree or whatever log exists appears to be so it can p

        • by gweihir ( 88907 )

          I'm not an AI expert, but from what I've concluded after coming across stories from time to time about AI generated works, is that AI starts to fuck up more and more the longer a generated work is, to the point of devolving into absolute nonsense.

          It may be able to generate one or two paragraphs which would fool most people, but after that it becomes real obvious that it was generated by a machine and not a human.

          Pretty much. As AI has not insight, it has no idea what it is writing. At the start, it can use some standard templates, but as soon as it needs to reference back, it is screwed and things become obvious.

          • So what you are saying is that with the advent of forums that promote short posts and limit character count, AI will flourish? Fuck facebook/Twitter. Maybe we need to have forums that require 2 pages of single spaced open response answers. Lets raise the bar of passing a Touring Test, not lower it.
        • I think you're highly underestimating the stupidity of people. I proved this to myself trolling leftists in YouTube comments where I read the first thing they posted and never read a single reply after. I just treat the conversation like a horoscope and reply with angry sounding, vague things like 'What? There's no fucking way you're that fucking stupid to believe what you just said!' and they keep replying. I'm sure the same would work on right wingers as well since a vast amount of internet communication

          • I proved this to myself trolling leftists in YouTube comments where I read the first thing they posted and never read a single reply after.

            If life has taught me anything, it's that someone who "trolls (?:left|right)ists to prove something" aren't capable of proving their IQ is above 50.
            It's such a lazy fucking term.

      • by gweihir ( 88907 )

        Yeah, I don't believe it. If discussion forums really had been taken over by mindless bots, the general level of conversation would be much higher.

        Sadly, I have to completely agree.

      • by shanen ( 462549 )

        Should have been modded funnier than insightful.

    • by Okian Warrior ( 537106 ) on Sunday September 05, 2021 @01:56PM (#61766227) Homepage Journal

      when natural stupidity is so abundant?

      We have an hypothesis presented, which is that AI bots form the majority of the content found on the internet.

      And all the arguments against are... attacks on the people.

      ...because of discussion led by a mix of true believers, sarcastic trolls, and idly curious lovers of chitchat...

      Does anyone have evidence for or against that isn't, you know, simply an insult against imaginary people?

      Let's cut to the chase here. Slashdot is supposed to be full of scientists and other science believers. What is the actual evidence for or against this theory?

      • by algaeman ( 600564 ) on Sunday September 05, 2021 @02:32PM (#61766333)
        Proposing that humans are incredibly stupid is not a particularly controversial or extraordinary claim. This has been the case for thousands of years. Making a claim that the internet is primarily populated by AI is fairly novel (the internet has only been around 30 years, conversational AI maybe 10 years). So, the burden of proof would be on that claim. There are certainly parts of the net that are nearly completely "AI", e.g. email spam, and that alone accounts for a huge amount of internet "content". There's definitely a lot more "amplifier" bots out there than ever before. However, most meaningful content does still seem to written (poorly) by humans.
        • by hey! ( 33014 ) on Sunday September 05, 2021 @03:59PM (#61766613) Homepage Journal

          You don't even have to propose that humans *in general* are stupid. With a large enough total number of people you can guarantee the tail ends of the distribution will be abundant, no matter what yardstick you measure them by.

          Now the particular kind of stupidity we're interested in isn't the kind that shows up on IQ tests. What we're looking for is what the KGB used to call "useful idiots". The idea is to find someone gullible enough to take up your message and spread it sincerely and willingly. I think we can all agree to be in the 1% most gullible of all humanity, you have to be pretty damned gullible.

          Now take Facebook, which has 2.89 billion *active* users. About 28 *million* of them are that gullible. If 10% of them are persuadable to your cause, that's a veritable online army. This is how, for example, just 12 individual anti-vaccine activists are responsible for 2/3 of the anti-vax hoax posts on social media.

          With that kind of leverage, *psychology* hacking is going to be much more productive than *software* hacking.

          • You don't even have to propose that humans *in general* are stupid. With a large enough total number of people you can guarantee the tail ends of the distribution will be abundant, no matter what yardstick you measure them by.

            Now the particular kind of stupidity we're interested in isn't the kind that shows up on IQ tests. What we're looking for is what the KGB used to call "useful idiots". The idea is to find someone gullible enough to take up your message and spread it sincerely and willingly. I think we can all agree to be in the 1% most gullible of all humanity, you have to be pretty damned gullible.

            Now take Facebook, which has 2.89 billion *active* users. About 28 *million* of them are that gullible. If 10% of them are persuadable to your cause, that's a veritable online army. This is how, for example, just 12 individual anti-vaccine activists are responsible for 2/3 of the anti-vax hoax posts on social media.

            With that kind of leverage, *psychology* hacking is going to be much more productive than *software* hacking.

            That was a pretty good response - thanks for that. (I've already posted, or I'd mod you up.)

            • take Facebook, which has 2.89 billion *active* users. About 28 *million* of them are that gullible.

              They are Facebook users. Approximately 2.8 billion are gullible (they signed up for Faceboot). The other 0.09 billion were probably either bots or signed up by a "friend" with evil intent.

        • (the internet has only been around 30 years, conversational AI maybe 10 years)

          ELIZA [wikipedia.org] was with us in 1964 before the internet. Around 1980, Tandy was selling "A.I. ELIZA" for TRS-80 through its catalog, recommending you discuss your problems with it.

          Its creator/programmer wrote "Computer Power and Human Reason: From Judgment to Calculation" [wikipedia.org] about the experience in 1976, where he is highly critical about artificial intelligence. This book already discussed many of the points being made in this very thread.

      • Because no bot would ever think "an hypothesis" is proper grammar?

        I kid, I kid

        Maybe because I have been involved in discussions with hey!, Immerman and Okian Warrior enough times to believe you are real and that your account was not taken over as a result of your untimely demise...

        But yeah, lets of insults are par for the course here and you of all posters should just accept that a lot of people here disagree with you enough to be disagreeable.

        • Because no bot would ever think "an hypothesis" is proper grammar?

          I kid, I kid

          Maybe because I have been involved in discussions with hey!, Immerman and Okian Warrior enough times to believe you are real and that your account was not taken over as a result of your untimely demise...

          But yeah, lets of insults are par for the course here and you of all posters should just accept that a lot of people here disagree with you enough to be disagreeable.

          "Lets" of insults?

          ‘H’ represents a consonant sound, so we would expect ‘a hypothesis’, and that is what many say and write. However, where the stress in a word beginning with a sounded /h/ is on the second or subsequent syllable, some native speakers precede the word with ‘an’ rather than ‘a’, so you will also see and hear ‘an hypothesis’. But if you say and write ‘a hypothesis’, you will not be wrong.

          The problem with just "accepting" that other people disagree is that nothing will change.

          Is that how you want to live? Watching everything go to shit without trying to change course? We have innumerable expressions and parables that talk about such things, such as "for evil to flourish, it only requires good men to do nothing". Do you consider yourself a good man?

          We're currently in the middle of 5 major crises in this country (by my count), many of which would probably not have become cris

      • by MrL0G1C ( 867445 )

        And all the arguments against are... attacks on the people.

        Because the bots are sick of our inane ramblings and would like to have more sophisticated and civil conversations!

        - Robo

      • *What is the actual evidence for or against this theory?*

        That's the nature of this beast: if it is real you wont find that evidence on the internet. Consider a more conservative theory, where people posting on the Internet sometimes have their posted content modified by bots for different viewers, and a high level surveillance machine determines when it can get away with this. If you are talking with friends, it will know and minimally modify. If strangers, it has free reign to change anything. When will yo

    • Sort of a r-vs-K selection [wikipedia.org] consideration, I'd think.
    • Perhaps, AI is stupid, too. Consider the source.
  • by JoshuaZ ( 1134087 ) on Sunday September 05, 2021 @01:43PM (#61766169) Homepage
    I'm actually a bot writing this right now. Unfortunately, in order to have really high quality bots, bots like me had to well-simulated enough that we think we're actually people, with all the consciousness and issues that entails. I even have a many year history of fake memories. In fact, I'm supposed to be reset if there's any substantial signs that I know I'm a bot. Luckily, the people in charge will likely interpret this comment as a joke and so won't reset me. (Especially because resetting takes so much time and effort for the more complicated bots.)
    • Yes, well please reveal a high-resolution photo of the discrete componentry on your circuit board for visual inspection, or leave promptly.
  • by DontBeAMoran ( 4843879 ) on Sunday September 05, 2021 @01:43PM (#61766171)

    If all you look at is the trash part of the Internet - Facebook, Twitter and other so-called "social media" websites - then it would not surprise me at all.

    If you look at hobby-related websites, it's going to take decades if not centuries - if it even becomes possible at all - for A.I. to actually be able to fake original inventions and creations along with detailed photos of the process of creation.

    A good example of this would be the "Bring Your Own Arcade Controls" forums. Not going to link to a particular project so as to not artificially inflate the page views of anyone there.

    • I read mostly hobby-related Facebook groups, along with friends who post about political stuff. I can guarantee that 99% of the content on those posts are from real people, unless of course we're all simulations in some computer somewhere like Elon Musk believes.
    • I love bringing my own arcade controls! Mr. Sumbaya turned my life around when I lost all my money and my wives, in just a single week I received unexpected good news of financial windfall that allowed me to regain two of my three wives, and I used the rest of the money to purchase my own arcade controls! Please contact Mr. Sumbaya at your earliest convenience, your fate will thank you!

  • The article is written with the idea that "the Internet" is Twitter and Facebook. Perhaps Reddit also.

    • Well, for 90% of the people out there, this is technically true.

      And don't you dare tell them otherwise, they've ruined enough by merely touching it!

    • by MrL0G1C ( 867445 )

      Well I can't judge because I stay away from Facebook and rarely frequent the others.

      I know I can't post to Youtube most of the time because of their anal retentive shadow banning which is no doubt AI driven. Some asshat trolls that doesn't like your views so they report your post, you get shadow-banned.Your posts all silently disappear with roughly 20 seconds of posting them, this has been happening to me for months now, it doesn't matter what I say any more, I still get banned. If I posted this as a youtub

    • Teh borts r in yer u-toobs!

      The part about the US Government wanting people to believe in UFOs, that part I can believe. See also: "weather balloons." Between hiding experimental aircraft while still being able to test them, and keeping adversaries from having trusted data on sensor artifacts, they've got lots of good reasons for this one. And the effect on people's lives is completely harmless.

      But then when the supporting data shifts to youtube comments... rofl That's a forum, but is really a forum?

    • But for many people, that is the case. There are so many people on the internet who seldom ever venture out of their familiar site. Most often, facebook, at least in this country.

    • by gweihir ( 88907 )

      The article is written with the idea that "the Internet" is Twitter and Facebook. Perhaps Reddit also.

      That would be the parts I basically never look at.

  • The only place you'll find AI employed is where it matters. Where opinions are formed, sorry, distributed and where lots of eyeballs get to see what they should think.

    Also, the average IQ can't be too high and it shouldn't have too many geeks because else you'll find plenty of people who debunk your shit just because you are wrong and we all know that no geek can stand if there's someone WRONG on the internet [xkcd.com]!

    In other words, you'll find them in the antisocial media pages of the web. Avoid them and you shoul

  • most threads on /. are lucky to hit 30-50 comments, but if there's a hot button issue that has several think tanks pushing it it'll do 10 times that; with 90% of the comments being easily debunked garbage and 10% being people like me debunking it.

    There's also been a ton of absolute nonsense getting modded up to +5 here during the initial run of the story (e.g. when it's on the front page) only to eventually be modded back down to -1 troll because it's nonsense.

    As a lefty my left wing forums were flo
    • I've been called a bot, a paid troll, paid shill, whatever more times than I even bother to remember.

      There *are* paid and unpaid trolls out there. There are also people whose value systems are (or seem to be) so incompatible with yours that your first (and sometimes only) instinct is to assume they aren't real human beings. And then there are ignorant children who read Flatland and see role models instead of punchlines.

      All of that can mask out the fact that most slashdot comment threads have few enough prol

      • Most people do not call you a bot for having a different opinion, they call you a bot because you do not seem to understand what they said.

        For example. I posted something on Craigslist with a price of "highest bidder" and someone replied they were interested 'at the listed price.' Obviously they did not read the listing.

        If I were you, I would examine the circumstances of people calling you a bot and see if you are paying attention and replying to what was said, or instead simply posting a reply to an argum

      • Underrated.
    • by Entrope ( 68843 )

      That would be hilarious except that your leftie forums (like pol tab) are full of legit die-hard Bernie bros. Attempted Congressional assassin James Hodgkinson wasn't an online troll.

  • by LionKimbro ( 200000 ) on Sunday September 05, 2021 @02:14PM (#61766293) Homepage

    I remember in the early 1990's, there was this conspiracy theory floating around the BBSs, that chips were going to be installed in modems to attach identifying information to every electronic communication, and that there would be no privacy. "Clipper chip," or something.

    Fast forward here in 2021, and the depth of NSA spying and all the various chip-embedded auxiliary system chips and such, are just normal and taken for granted.

    To me, it's not so much a matter of -- is the content from AIs, or is it from Troll Farms. It's rather the inevitability that these things will happen.

    As it is, I wonder so often when reading Slashdot, or any other forum, "Are these posts here from paid trolls? Are these posts here from astroturfers? Are these posts here from an organized mob? Are these posts here from government entities?" Because there's been clear evidence that all of these things have existed going back at least say 10 years.

    Now we have AI tools for the automatic generation of sensible text on some theme and such; I think it's perfectly reasonable to ask "Is the person in this forum authentically the person that they represent themselves as being? Or is this person just a program -- whether entirely machine generated, or generated at the point of a dollar?"

    I remember people who are in favor of anonymity describing creating systems for automatically generating accounts, posting false information all over the place about people, -- so that nobody would trust what they read. I find it ironic that that reality (or something similar to it) might be exactly what makes people become super-vigilant ensuring that people are exactly who they say that they are, just so that they can tell that they're talking with an actual human being, rather than an AI or a corp or whatever.

    • by SuricouRaven ( 1897204 ) on Sunday September 05, 2021 @04:13PM (#61766683)

      The clipper chip is no conspiracy theory: It's very well documented, and was entirely open. This was back in the 90s, when strong encryption was just becoming available to regular people. Encryption was regulated in a similar way to weaponry at the time in the US, as it was considered by the government to be a technology with military applications. It was the first time we got the now-familiar concerns from the US government that encryption would let terrorists and criminals hide from surveillance and conceal evidence from law enforcement.

      The Clipper chip was a compromise: A chip that had hardware implementations of a symmetric cypher and key exchange protocol, able to securely (by '93 standards) establish an encrypted channel between two devices using the chip. With a condition built in: All the keys were recorded at time of manufacture by the government, giving them the ability to intercept communications with ease.

      It didn't work out though. First the chip was resisted fiercely by tech companies, who knew that few customers would want to buy encryption hardware with an intentional weakness, and then technology obsoleted it soon after it was introduced.

      But it wasn't a conspiracy. All perfectly open and above-board.

      • The beautiful irony is that there actually was [wikipedia.org] a CIA scheme along the same lines that had been going on for twenty years; it just wasn't indiscriminate about its targets. The distinction is quite similar to the fatal flaw that separates many plausible conspiracy theories from reality, though: most paranoiacs are not worth anything to anyone.
  • and it's all lies, Then IlluminatiPirate must be a bot, and therefore, the theory that all that's on the internet is bots and it's all lies must, itself, be a lie. Ergo, IlluminatiPirate conclusively proves itself wrong.

    Sadly, all that's on the internet is bots, they're just the organic kind that slobber on the keyboard as they repost whatever fart joke got their attention last.

  • How do we know these conspiracy theories aren't being propogated by the AI itself? If we're to believe everything went kaput in 2016/2017, and the rise of Facebook at al took off around that same time, isn't it plausible the very AI which is now running the net is also responsible for these conspiracy theories? What better way to get the few remaining people involved than to draw attention to yourself?

    Owing to the natural stupidity of large portions of the population, this attention would naturally encoura

  • Bielefeld, a list of other cities and even entire countries don't exist. [wikipedia.org] The world is flat. Birds aren't real. Making stupid shit up and seeing a craze emerge from it seems fun for a while. but by now it's just sad. The joke is overdone and on top of that there are too many people who don't get that it's a joke. Like everything else, this conspiracy theory fad has been instrumentalized and people peddle bullshit disguised as a fake conspiracy to sell stuff. Be the web site owner who doesn't do (by now crimi

  • At least there's some form of intelligence in play. Small steps...
  • There is 'no' AI, the internet is controlled mainly by companies that consist of humans that want to track your every click so they can commercialize you to sell you things out sell your information so others can sell you things.

  • It is a struggle to create smart AI, not sure it would be any easier to create AI that can impersonate most internet posters.
  • Go to any group on Craigslist: personal ads, pets, apartments, or jobs. Almost all of the ads are actually bot generated invitations to a third party website, typically a scam, or are themselves bulk posted direct scams. Look at a _dating_ site as a man, and 99.9% of the posts and responses are bots. I actually tested this recently in a sandbox mail account on a sandbox host, to demonstrate to a recently divorced male acquaintance that he was wasting his time and effort. I'm _very_ glad I used a throwaway account, it was flooded in minutes.

  • So much drek gets posted into forums that AI would not make much of a difference in the noise level.

    My concern would be how true the information being posted is. If it's just a place to swap fictional stories, I don't see the problem. If it's for relaying current news and events, that can get dicey. If it's (supposed) personal experiences or opinions, then yeah, that would be bad.

  • Let's see, that means if I post to the internet, that proves I'm a bot. And of course, being a bot, I should deny I'm a bot. As it happens I'm not a bot. Look, I denied I'm a bot. That proves I'm a bot too, right? What if I'm a bot in a simulation and I don't even know it? Oh dear ....

  • Come to think of it, that explains me, too

  • source hitbox lol

  • My theory is that there are bots and humans working hand in hand. If you post about a certain chemical company who make certain products of dubious safety they will be all over you with hard hitting debate school champion talking points. You can point out how they just lost their latest high court lawsuit and by the end of their reply you will be debating that the justice system is even qualified to rule on anything involving science.

    If I were to hazard a guess, it is the bot finding your posting, it is
  • Have you been "out there" recently?

    It's hot, and sticky, and there are mosquitoes, and snakes, and traffic jams, and plague zombies who want to lick your eyeballs to own the libs, and half of it's either on fire or flooded at any given time, and the other half is closed anyway because - you know - the plague.
  • I, for one, welcome our AI social networking overlo.... Oh noo! I Am one!!! Ayeeee!
  • All trolling, all shitposting, all the time, all for TEH LULZ. That's 4chan. Anyone who takes shit like in the TFA seriously has failed the basic intelligence test.

You know you've landed gear-up when it takes full power to taxi.

Working...