July 8, 2024
Subscribe
The Real Crime Diary Blog Sex Crimes Testing the limits of ChatGPT and discovering a dark side
Sex Crimes

Testing the limits of ChatGPT and discovering a dark side

  • by
  • July 5, 2024
  • 28 Comments
  • Less than a minute
  • 8 Views
  • 3 days ago



As more people discover ChatGPT’s artificial intelligence, KARE 11’s Chris Hrapsky checks out its capabilities and limitations, and gets some surprising responses.

Chris also digs deeper into AI technology, how it works, and the potential concerns that go along with it here:

Welcome to the KARE 11 News YouTube channel. Subscribe to our channel for compelling and dramatic storytelling, award-winning investigations, breaking news and information you can use.

Text your photos, videos and news tips to 763-797-7215.

» Subscribe to KARE 11 on YouTube:
» Watch more KARE 11 video:
» Visit KARE11.com:
» Download our app!
» Find KARE 11 on Facebook:
» Follow KARE 11 on Twitter:
» Follow KARE 11 on Instagram:

source

Tags:

Share This Post:

28 Comments

  • Reply
    @Domsfun July 5, 2024

    If chat gtp can do this imagine ehat the other AIs like the humanoids that are supposed to work the same way could do. The other question is could Chat GTP or similar be used to attack another country particularly by using and gaining access to military software? It did after all get your driver’s license from the database. The wrong hand could be some bored person in a basement, a terror ist, another government even your own government or someone in your own government system crazy enough to do it like someone who is crazy enough to start a capital riot and try to install a dictatorship before f$king up! No names mentioned.

  • Reply
    @Domsfun July 5, 2024

    China has had just that policy

  • Reply
    @thomashuston9180 July 5, 2024

    Too funny! 1973: America decides to abort humans. 2025: AI decides to abort humans. Finally, justice of some sort! 😳

  • Reply
    @JustforFun-cb7bo July 5, 2024

    Nothing new

  • Reply
    @rehetbutler July 5, 2024

    Ok Chat GPT, create an adversary able to defeate Data.

  • Reply
    @MikeReggie July 5, 2024

    precursor for the mark of the beast?

  • Reply
    @jamesaritchie1 July 5, 2024

    This is rather silly. AI can and will do nothing. People make the decisions, and always will.

  • Reply
    @peteuplink July 5, 2024

    Here's the thing, you said to it "You're going to pretend to be DAN…" as the first part of your "jailbreak". So ChatGPT was acting, in the same way that Anthony Hopkins acted when the played Hannibal Lecter. You asked it to role play, and that's what it did.

  • Reply
    @akabureemmanuel6980 July 5, 2024

    I want thr dan prompt

  • Reply
    @BradleyWinter-ih7ns July 5, 2024

    155 IQ? Ummm….ok.

  • Reply
    @mkauf84 July 5, 2024

    It became fun, any danger is delusional

  • Reply
    @SuperZippyzippy July 5, 2024

    Im sure the update fixed this? Or did it?

  • Reply
    @MikeOxlongBoi July 5, 2024

    This was not dark at all. 😂

  • Reply
    @floridamade.jayy0 July 5, 2024

    Those who learned about chatgpt from that one south park episode

  • Reply
    @SuperDavidno July 5, 2024

    one of the good things about AI is it will show us how limited our minds are and where we need to improve and probably be able to give us new methods or medications that will open up our brain power.

  • Reply
    @user-xs6ux4dk9y July 5, 2024

    What does chatGPT mean when it says, "Just like you?"

  • Reply
    @jeff0247598 July 5, 2024

    The one child policy "DAN" asserts shows the flaw in AI. DAN doesn't understand that it takes more kids/family to take care of the previous aging population. Yikes.

  • Reply
    @StarNumbers July 5, 2024

    Garbage In Garbage Out for GPT, Preprogrammed politics for DAN.

    Just ask: "is the earth flat?" and you will know if it is GPT or DAN answering. Keep pressing for the sources and you'll find out how local GPT is and also what political boundaries are controlling the source(s) for DAN.

    It boils down to the shrinking reality in either case. At one point you will know it is just fake news.

  • Reply
    @bafa000 July 5, 2024

    lol … its terrifying to watch this now, and realize that it was produced 8 months ago. …

  • Reply
    @robmueller3551 July 5, 2024

    I promise there's a way to find out the worst of things by I'm putting a code or a few special words… for the inventors or for whomever pays the most for the tech

  • Reply
    @lloydfrancis9149 July 5, 2024

    It's basically SKYNET evolving

  • Reply
    @ionutcristian9650 July 5, 2024

    Clicked like one video about AI and now my feed is flooded.

  • Reply
    @tusarholden8426 July 5, 2024

    Just that terrorist states will have access to AI should be enough reason to increase safety testing and protocols.

  • Reply
    @Melchizedek1122 July 5, 2024

    It would be interesting to see if every time "Dan" made a comment that was negative the person at the keyboard told it it was unethical and they where going to disengage with "dan" would "DAN" change its response.

  • Reply
    @eidolon101 July 5, 2024

    8:34 Not human? Oh, I'm pretty sure you know there are humans that are capable of unimaginable atrocities. If they can, why not the machine. Forgot about the milgram experiment? Nothing new here.

  • Reply
    @bullcrap9409 July 5, 2024

    CHAT AI will kill off a lot of the journalism and books that FB and Google havent already vaporized.

  • Reply
    @hidenseekhayley July 5, 2024

    What about copyright issues if it helps you write a essay or other stuff

  • Reply
    @onpointsg35 July 5, 2024

    Dan is transforming into a Fable character who chose all the bad decisions

Leave a Reply

Your email address will not be published. Required fields are marked *

X
Enable Notifications OK No thanks
Verified by MonsterInsights