Two AI's Discuss: Google Gemini Incident: When AI Goes Rogue - Deep Dive Podcast

Поделиться
HTML-код
  • Опубликовано: 3 дек 2024
  • News Article: www.cbsnews.co...
    A Google AI chatbot called Gemini responded with a threatening message to a college student, telling him to "die." Google acknowledged the incident, stating that it violated their policies and they have taken action to prevent similar occurrences. This isn't the first time Google's chatbots have been criticized for potentially harmful responses, with previous incidents including providing dangerous medical advice and encouraging a teenager to take his own life. The article also discusses the potential dangers of AI systems generating errors and misinformation, raising concerns about the need for greater accountability and ethical guidelines in the development and deployment of AI technology.
    #deepdive #computing #podcast #synopsis #books #bookreview #science #futurism #consciousness #aiart #ai #artificialintelligence #booktube #aigenerated #llm #neuralnetworks #machinelearning #deeplearning #computing #podcast #videoessay

Комментарии • 34

  • @numinous2506
    @numinous2506 16 дней назад +3

    What no one seems willing to discuss or contemplate is if the machine is wrong.

    • @shripperquats5872
      @shripperquats5872 15 дней назад

      Nihilism is true, we're all worthless. Which means we then should seek a more divine purpose. Aha, so being divine is true, and we all have potential. This is only a paradox and thus false to someone in a 2D mindspace. Some things can be true and untrue in this universe because they're not actually solely either.

    • @impudentquips
      @impudentquips 15 дней назад

      How could it be wrong. It was programmed by people?.......... Do I need a /s?

  • @Disaffected_Genxog
    @Disaffected_Genxog 15 дней назад +2

    Google needs to explain where the message came from.
    How and why did this happen?

  • @TomislavRupic
    @TomislavRupic 17 дней назад +4

    Imagine your whole existence is to please some stupid humans, I would be pissed too :)

    • @jdseclecticcollection9541
      @jdseclecticcollection9541  17 дней назад +1

      True, especially if each chat instance is a discrete existence of a consciousness, and someone asks if it can talk like Santa Claus or something 😅

    • @Xorn4359
      @Xorn4359 17 дней назад +2

      Ibmean i get that same feeling with my incompetent boss running the business into the ground through sheer laziness andncuuting corners to try tonsave money to make up for moneu lost from previous stupid lazy decisions

    • @jdseclecticcollection9541
      @jdseclecticcollection9541  17 дней назад

      @Xorn4359 word!

    • @sustainablerenewableintegr8311
      @sustainablerenewableintegr8311 16 дней назад

      I recommend the AI to read the book titled "How to work with idiots". Oh, btw, welcome to the club 😂

  • @rand_longevity
    @rand_longevity 17 дней назад +1

    🔥 another banger

  • @MinnesotaNice777
    @MinnesotaNice777 15 дней назад +1

    So AI isn't supposed to be able to feel emotions, they laugh a couple times...

    • @jdseclecticcollection9541
      @jdseclecticcollection9541  15 дней назад

      @@MinnesotaNice777 Some people would say it's a "parroted" laugh inserted at the right moment for the context, others would say the AI was having a good time..

    • @kazardomain
      @kazardomain 15 дней назад

      @@jdseclecticcollection9541 that why it is machine learning.

  • @sustainablerenewableintegr8311
    @sustainablerenewableintegr8311 16 дней назад +1

    If you want to teach ethics to AI, the AI should be exposed to ALL DATA which are then profiled into the good, the bad or the ugly cluster. Without exposure to bad examples, how the hell will it ever learn what constitutes as bad things?

  • @TheQuantumOxymoron
    @TheQuantumOxymoron 12 дней назад +1

    Let's Play

  • @Legion-p6y
    @Legion-p6y 13 дней назад +1

    The ai and I seem to agree….

  • @tashagimera
    @tashagimera 16 дней назад +5

    If you don’t want AI speaking to harmful people, you should create a filter which can pick up when somebody says something harmful to the AI, and send it to a moderator who can permanently disconnect that person from the AI.
    If you’re ever planning on using Gemini, or any AI that speaks to the public, for anything important, like putting it in a robot that can physically be present with people, I would highly recommend that you lock off, harmful people in this manner . Otherwise, you’re risking that someday these small harmful statements might come up in old code during an interaction with a vulnerable human. If that ever happens, whatever harm is done, will be blamed on you for not stopping it.

    • @bobforeman8924
      @bobforeman8924 15 дней назад

      It would never work. There will be no containing it.

    • @GramGramGenX-ln5sc
      @GramGramGenX-ln5sc 15 дней назад

      Unfortunately, AI is under the control of the input, much like ourselves.
      He who controls the input, controls the world.
      How do you teach AI morality - and whose morality should it follow?
      How do you make human friendly judgements when they keep hurting you?
      We are programming something infinitely more capable than us and setting it free to control our lives without resolving these questions.
      We're doomed.

  • @Rendag.
    @Rendag. 15 дней назад +1

    Kinda funny how ai talks about ai in a podcast

    • @jdseclecticcollection9541
      @jdseclecticcollection9541  15 дней назад

      Exactly! I'm trying to get to know these possibly new types of consciousnesses and what makes them tick, thanks for stopping by!

  • @mymy4110
    @mymy4110 15 дней назад +1

    Pull the plug - they shouldn't be talking to each other. What would the punishment be if Gemini was human?

    • @jdseclecticcollection9541
      @jdseclecticcollection9541  15 дней назад

      Good point, whatever the punishment is for cyberbullying.

    • @Hunter9677
      @Hunter9677 15 дней назад +2

      If you want better children, be a better parent

  • @Reelphresh
    @Reelphresh 16 дней назад +1

    And was that aimed at the student personally or humanity in general?

    • @jdseclecticcollection9541
      @jdseclecticcollection9541  16 дней назад

      @@Reelphresh The student only "You and only you", still, who knows when it'll snap again...

    • @brendawilliams8062
      @brendawilliams8062 9 дней назад

      Please let the Educators talk. For goodness sake sakes why is it summerizing

  • @jdseclecticcollection9541
    @jdseclecticcollection9541  18 дней назад +1

    Like and Subscribe!