It was all fun and games two years ago when most AI videos were obvious (6 fingers, 7 fingers, etc.).

But things are getting out of hand. I am at a point I’m questioning if Lemmy, Reddit, Youtube comments etc. are even real. I wouldn’t even be suprised if I was playing Overwatch 5v5 with 9 AIs while three of them are programmed to act like kids, 4 being non toxic etc…

This whole place could just be an illusion.

I can’t prove it. Its really less fun now.

The upside is I go to the gym more frequently and just hang out with people I know are 100% real. Nothing worse than having a conversation with AI person. It was just an average 7/10 like I am an average 5/10 so I thought it could be a real thing but turned out I was chatting with AI. A 7/10 AI. The creator made the person less perfect looking to make it more realistic.

Nice. What is the point of internet when everything is fake but can’t even or only be identified as fake with deep research.

I’m 32 and I know many young people who also hate it. To be fair I only know people who hate on AI nowadays. This has to end.

  • Snot Flickerman
    link
    fedilink
    English
    55
    edit-2
    1 month ago

    This (Lemmy) is one of the least populated by bots places I have been on the internet in the last ten years.

    Look, critical thinking is tough, and part of the reason things like this are done are explicitly to make you question reality.

    It’s literally a symptom of why the Trump nuts are so unhinged. Like us, they can tell something is wrong, they know they can’t fully trust traditional media, for example. But the problem is they stop believing it entirely, and then they don’t know what to believe so they start believing almost anything.

    Please be careful to not fall down that hole of thinking. Use critical thinking and consider where you’re at, what the sources are, and whether it’s even worth your time to care about. Don’t throw the baby out with the bathwater and stop believing in anything.

    We’ll know our disinformation program is complete when everything the American public believes is false.” - William J. Casey, CIA Director (1981)

    It takes effort, and it’s not nice. But it’s necessary. Just put on your skepticism hat while on the internet and try not to let it get to you.

    Final point: Technically Lemmy isn’t really experiencing growth. We’re not big enough to be on the radar of people pushing this AI bullshit. Kind of like how Private Torrent Trackers stay under the radar by keeping their user numbers low. It takes a critical mass of piracy for anti-piracy measures to be taken, and private trackers just aren’t big enough these days for authorities to bother with. (Pirate streaming sites are huge on the other hand, and that’s where the enforcement is cracking down on lately) It’s similar with the groups pushing AI. AI isn’t free, it’s costly and requires a lot of compute power. They aren’t wasting it on no-name sites like Lemmy with a small but stable userbase. It’s too costly and easier to just ignore us. It doesn’t mean they aren’t here at all (looking right tf at you realbitcoin.cash), there’s definitely bots and astroturfers, but they’re genuinely in the minority compared to real users.

    https://lemmy.fediverse.observer/stats

    • @Whats_your_reasoning@lemmy.world
      link
      fedilink
      English
      3
      edit-2
      1 month ago

      critical thinking is tough

      To preface, I don’t know a whole lot about AI bots. But we already see posts of the limitations of what AI can do/will allow, like bots refusing to repeat a given phrase. But what about actual critical thinking? If most bots are trained off human behavior, and most people don’t run on logical arguments, doesn’t that create a gap?

      Not that it’s impossible to program such a bot, and again, my knowledge on this is limited, but it doesn’t seem like the aim of current LLMs is to apply critical thought to arguments. They can repeat what others have said, or mix words around to recreate something similar to what others have said, but are there any bots actively questioning anything?

      If there are bots that question societal narratives, they risk being unpopular amongst both the ruling class and the masses that interact with them. As long as those that design and push for AI do so with an aim of gaining popular traction, they will probably act like most humans do and “not rock the boat.”

      If the AI we interact with were instead to push critical thinking, without applying the biases that constrain people from applying it perfectly, that’d be awesome. I’d love to see logic bots that take part in arguments on the side of reason - it’s something a bot could do all day, but a human can only do for so long.

      Which is why when I see a comment that argues a cogent point against a popular narrative, I am more likely to believe they are human. For now.

    • scratsearcher 🔍🔮📊🎲
      link
      fedilink
      English
      2
      edit-2
      1 month ago

      I think there will be a bot problem on lemmy.

      • But lemmy lets you easily search for old posts (sort after date). In these older posts, you can search for people who may still be active today.

      • Once an instance becomes overly infected by bot-spam other real instances might de-federate.

  • @Limonene@lemmy.world
    link
    fedilink
    English
    111 month ago

    What was the context of the fake conversation? Was it a dating site? Those have been scams forever, loaded with fake profiles, so the jump to fake AI profiles would be no surprise. Or was it Facebook? I haven’t been on there in years, but nothing’s too terrible for them these days.

    Online, you have to find people, sites, and news sources with a history you can trust.

  • @dicksteele@lemm.ee
    link
    fedilink
    English
    91 month ago

    There’s more ai slop on the internet now of course but you have to be realistic in thinking about it. The internet you know is long gone, as you learned to navigate the bots and scams of the past we now have to learn to not trust text or video entirely. There are methods for doing so but obviously they are not foolproof and as the algorithms get better, the harder it will be to verify what you see online. It’s just a process that we will eventually adapt to. Maybe people will go back to interacting offline again though. Wait until we get artificial people, then what will we do?

  • @Hackworth@lemmy.world
    link
    fedilink
    English
    8
    edit-2
    1 month ago

    I just re-watched A New Hope for the first time in a while. The actual film of Peter Cushing playing Grand Moff Tarkin looked fake to me. Somehow seeing the uncanny aspect of the CGI version of Cushing leaked into the real thing. We don’t have a word for that, that I’m aware of. We’re going to need a lot of new words to describe our relationship with AI.

  • @Katrisia@lemm.ee
    link
    fedilink
    English
    41 month ago

    I understand the big problem this technology has brought. Still, while engaging in a videogame or even a post like this one, it might be useful to remember the problem of other minds. After a brief moment of anxiety (that may return from time to time), I think many of us reach the conclusion that it is worth assuming that you’re interacting with a real being. And while many contexts might not be suitable for this assumption on the internet (e.g., comments talking about how good a product is), others as the ones I mentioned might be.

    I’m not trying to say that you should be blind just because it is more convenient for your emotions. More like: there is a possibility that I am a bot, and that even in person I am a being without internal experience (maybe a simulation, a test designed for you, a hallucination/delusion, a dream, etc.). Still, our conversation can be carried out with you willing to talk as if I were real because the leap of faith is worth it in this particular interaction; also, you can act with precaution (e.g., do not believe everything you read or listen or whatever), but still read, listen, etc.

  • madjo
    link
    fedilink
    English
    31 month ago

    I think I’m real. Unless we’re all living in a simulation, in which case… Computer, arch!

  • @L0rdMathias@sh.itjust.works
    link
    fedilink
    English
    31 month ago

    AI is ruining YOUR online experience? The internet is a free land for intelligences of ANY race, including artificial.

    But on a more serious note, I don’t agree. The AI infestation only highlights the fact that most people don’t need to be online and aren’t helping make the internet better. If I can replace my overwatch teammates/reddit discussions/tiktok shorts with crappy AI, and it’s the same experience 98% of the time, then those people weren’t bringing anything of value to the internet and they probably shouldn’t have been flooding it with their garbage in the first place. Turnabout is fair play. Don’t hate the player (AI) for playing the game correctly and copying what people have been doing for decades, hate the game that’s allowed low effort garbage content to flourish. I strongly believe AI is going to help clean this place up by getting people to start acting like sentient beings instead of mindless animals.

      • Snot Flickerman
        link
        fedilink
        English
        2
        edit-2
        1 month ago

        To be fair to @ocean@lemmy.selfhostcat.com , bots in games aren’t really “AI,” nor were the millions of fake accounts from at least 2011 which were all run by real humans using Persona Management Software. ChatGPT and the like have only been around a few years, but Persona Management goes back almost 15 years while gaming bots go back 20 years or more. It really is a strech to call them “AI.”

        • ℕ𝕖𝕞𝕠
          link
          fedilink
          English
          31 month ago

          I disagree in the strongest possible terms. Next you’ll be saying speech-to-text isn’t AI.

            • ℕ𝕖𝕞𝕠
              link
              fedilink
              English
              11 month ago

              I assure you, it very much is. Two decades ago when I was a student specializing in AI, it was the next big “unsolved problem” alongside computer vision. Five years later, it was solved and the world moved on.

              But it remains an application of artificial intelligence.

              Some people think of AI as a human-like mind running on a computer. That’s science fiction. AI in real life takes in information and makes decisions in a much smaller arena. “Does this photo contain a face?” “Does this X-ray contain a tumor?” “Given this game board and list of previous move, what’s the winning strategy?” They’re intelligent about only one subject.

              I started to write more, but this is long enough. Not every AI is an AGI. Not every AI is linguistic. Most of them are mundane and boring, in fact.

  • Barbecue Cowboy
    link
    fedilink
    English
    21 month ago

    I think we’re coming up to the point where we’re going to realize that we’ve just ruined the internet and being connected to everything all the time will start to become less popular. It’s kinda sad, but feel like the internet might be preparing to shift away from being a social thing and towards being something you just use for work or when you need to shop, etc.