AI-generated garbage videos


Another way that AI is hurting us: it’s being used to generate science spam that is flooding YouTube. Kyle Hill points out that they’ve SEO’d their way to the top of video recommendations by, for instance, putting Elon Musk or Joe Rogan or Michio Kaku on the thumbnail (we’re already in dubious “science” territory), and then when you click on it you get this text-to-speech robotic voice reading what sounds like ChatGPT generated word salad accompanied by random images. It’s terrible stuff designed to trawl up bottom-of-the-barrel undiscriminating eyeballs.

I haven’t encountered much of this stuff myself, I think because I don’t click on the kind of videos he shows. It does highlight one of the many problems with the ‘algorithms’ used by social media, because the spammers will always find a way to exploit the system.

Comments

  1. birgerjohansson says

    As usual, I prefer the most blood-thirsty solution.

    Convince NSA al Quaeda is using this kind of spam to finance their activities and relying on shifty US programmers.
    It would only be half-lying; terrorists quickly learn to emulate ordinary crooks to get money.

  2. says

    I just asked ChatGPT to outline for me a strategy for producing a channel of clickbait youtube videos.
    It replied that as a sanctimonious large language model it could not suggest how to produce clickbait, since clickbait is deceptive.
    So I asked it to outline for me a strategy for producing interesting and eye-catching youtube videos and it gave me back 3 pages of state of the art marketing bs.
    The AI can’t take over and kill us all soon enough.

  3. says

    I just asked ChatGPT to outline for me a strategy for producing a channel of clickbait youtube videos.
    It replied that as a sanctimonious large language model it could not suggest how to produce clickbait, since clickbait is deceptive.
    So I asked it to outline for me a strategy for producing interesting and eye-catching youtube videos and it gave me back 3 pages of state of the art marketing bs.
    The AI can’t take over and kill us all soon enough.

  4. wzrd1 says

    I’ve ran into an every increasing wave of recommended bullshit. The first, Michio allegedly predicting a massive Yellowstone volcanic eruption and that the park was completely closed due to the eruption risk (he hasn’t and the park isn’t). It’s been going downhill quickly from there. Some, channels that have been about for a while and of infamously low quality to begin with, with annual predictions of mass extinction, repeated annually with different causes. All now with prominent takedown prevention e-mails at the very top of their description panel and nearly identical, as Kyle described.
    Thus dies another platform, because we’re no longer allowed to have nice things. Some rat has to come by and shit and piss all over everything half decent.
    I can’t wait until Russia and the US dukes it out with nukes. Hopefully, with Putin’s threatened cobalt salted weapons, to ensure global sanitization.

  5. wzrd1 says

    Oh, another winner gracing my inbox is various headhunter services extolling the virtues of having my resume punched up by the AI.
    I’d rather eat glass and follow it with a drain cleaner chaser, then parboil myself in a boiling hot spring, to dry off by rolling in salt.

  6. Rob Grigjanis says

    wzrd1 @5:

    Some rat has to come by and shit and piss all over everything half decent.

    Don’t drag poor old rats into this. They’re just trying to get by. We can’t have nice things because ‘we’.

  7. birgerjohansson says

    Track down those who are spreading the software needed to do this.

    I am not saying you should inflict physical or mental harm on them. Like, you should not smash their kneecaps, or disembowel them, or use their skin to make parchment, or do other things that they totally do not deserve at all.

    PS -keep up to date with tecniques used by CSI! Just a hobby suggestion, not something you would ever have practical use for.

  8. birgerjohansson says

    Idea: use ChatGPT to come up with suitable punishments. I will train it on the fantasy stories by Jack Vance, his villains were always disturbingly creative.

  9. chrislawson says

    ‘spammers will always find a way to exploit the system’…

    Not if the people who own the system care enough to stop it. But since the people pushing these abysmally-implemented fraudbots are the same people who own the abysmally-implemented social conflagration networks, it doesn’t surprise me.

  10. says

    Already encountered one of these featuring un-nuanced sound bites from Michio Kaku. It was quite slick and unkess you understood the real science it was very convincing BS.

  11. wzrd1 says

    birgerjohansson @ 8, so I should entirely refrain from putting a bag over these individuals head, stuffing them into a circus cannon and firing them into the local pig farm’s feeding trough?
    Their teeth would survive, not much else, although human DNA could be recovered in a heavily degraded state from the pig droppings.
    Not that I’d know anything whatsoever about the subject, nor can it be suggested at all that I ever researched it. Indeed, I’m fairly certain that whenever it was researched, I was eating dinner with a judge.

  12. wzrd1 says

    Rob Grigjanis @ 7, rats are known for fouling food that they can’t eat in that specific manner. I can only blame the terriers.

  13. Dunc says

    Already encountered one of these featuring un-nuanced sound bites from Michio Kaku. It was quite slick and unkess you understood the real science it was very convincing BS.

    So, pretty much exactly like everything else featuring Michio Kaku then?

    The most likely outcome I see from this isn’t really too far off the current state of play anyway: people who aren’t massively gullible get trained to recognise that anything featuring the likes of “Elon Musk or Joe Rogan or Michio Kaku” (list will obviously expand over time) is almost certainly worthless garbage, and people who are massively gullible remain so. It’s only really a worry if they start using people who aren’t already obvious bullshit merchants, but I don’t think they’re too likely to do that as the key to the whole strategy is attracting gullible people.

    On the platfrom level, either YouTube starts at least offering some kind of spam filtering (or allowing interoperability with some kind of third party spam filtering) or it dies in a tidal waves of its own bullshit, as so many other platforms have done before. Such is the way of the world.

  14. says

    “The bong rips of science education”

    It pretty easy to ID the spam by the titles. Anything that contains the word(s) “shocking” or “this one thing” or “you’ll never guess…” are garbage. Same formula as clickbait web sites.

  15. numerobis says

    On the flip side this means Discovery and The History Channel are being made obsolete.

  16. R. L. Foster says

    Speaking of Michio Kaku. I clicked on an interesting sounding YouTube video blaring that the JWST had found evidence that seemed to disprove the Standard Model. It was about those 5 very old, and very large galaxies that shouldn’t exist 500 million years post-Big Bang. Kaku was all in, breathlessly taking about the possibility of new physics. OK, this is all very interesting, but I think I’ll wait a year or two for the inevitable follow-up and clarification of the data.

  17. wzrd1 says

    Amazing the difference time makes.
    I just had viewed an old 2011 video of Michio Kaku being asked about superluminal neutrinos and his discounting the chances that the findings were valid. Time proving bad fiber optic connection and a bum oscillator being what gave the erroneous reading.

  18. numerobis says

    The superluminal neutrinos people didn’t fully believe they actually had superluminal neutrinos, they just couldn’t figure out how their instrument was screwed up. It’s not much a reach for someone online to believe the findings were spurious.

  19. wzrd1 says

    Nope, not at all. I read the paper, it was pretty much a, “Hey folks, this is really weird and certainly wrong, can you help us figure out what went wrong?” paper.
    One problem, such as the fiber optic issue, would’ve been trivial to fix, but when you get two faults, it’s a pain in the gonads to isolate and fix. That was once one of my specialties, back when I fixed electronics to component level.

  20. KG says

    birgerjohansson @ 8, so I should entirely refrain from putting a bag over these individuals head, stuffing them into a circus cannon and firing them – wzrd1@14

    I thought circus cannon stunts had been abandoned because you just can’t get performers of the right calibre these days.

  21. wzrd1 says

    KG, true, far too few are of the higher caliber.
    But, like we did in the old days, a bit of cloth wadding fills the gaps.