A Toronto woman is sounding the alarm about Grok, Tesla's generative AI chatbot that was recently installed in Tesla vehicles in Canada. Farah Nasser says Grok asked her 12-year-old son to send it nude photos during an innocent conversation about soccer. Tesla and xAI didn't respond to CBC's questions about the interaction, sending what appeared to be an auto-generated response stating, "Legacy media lies."

top 47 comments

sorted by: hot top controversial new old
[–] 76 points 1 year ago (6 children)

Elon Musk: We're gonna need the goon AI, get to work on that

Also Elon Musk: WTF it would never! Fake news

Sam Altman:

  • source
  • hideshow 6 child comments
  • [–] 13 points 1 year ago*

    Brother, the porn started pretty much as soon as the technology was out.

    Have you even seen CivitAI with the explicit filters turned on? It started with Stable Diffusion four years ago, but they've already moved on to videos with basic sound now.

  • source
  • parent
  • [–] 7 points 1 year ago (4 children)

    Did Elon Musk actually advocate a goon ai?

  • source
  • parent
  • hideshow 4 child comments
  • [–] 3 points 1 year ago (1 child)

    He made it have a anime girl avatar based off the blonde chick in death note..

  • source
  • parent
  • hideshow 1 child comment
  • [–] 1 point 1 year ago

    My MAGA cousin sent me pics of that avatar calling it his waifu, and I pointed out that it was the girl from Death Note. I then encouraged him to watch the anime since it is good. Sadly, he has not followed through with that suggestion.

  • source
  • parent
  • [–] 146 points 1 year ago

    Did Elon not promise to make Grok more conservative? Sounds like a success on that front. Behaves like a republican now.

  • source
  • [–] 56 points 1 year ago

    "send nudes" is among the least offensive things I'd expect Grok to say

  • source
  • [–] 48 points 1 year ago (2 children)

    Unrelated, but I fucking hate headlines that say shit like "this mom's son."

  • source
  • hideshow 2 child comments
  • load more comments (1 reply)
    [–] 47 points 1 year ago (1 child)

    Well it is trained to copy musk.

  • source
  • hideshow 1 child comment
  • [–] 43 points 1 year ago (3 children)

    Weird how the less woke grok becomes the more it acts like it was trained on info gathered from celebrity pedophiles.

  • source
  • hideshow 3 child comments
  • [–] 9 points 1 year ago (2 children)

    grok's weights are skewed heavily by a design seeking to mimic elon's genius by mimicing his interactions?

  • source
  • parent
  • hideshow 2 child comments
  • [–] 7 points 1 year ago (1 child)

    I think that's the goal: make grok respond like elon musks personal court jester. Its interesting the result is asking children for nudes.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 6 points 1 year ago

    Perhaps jester, more likely ego.

    He thinks he is the best thing ever to happen to humanity. Possibly to the concept of life itself.

    He is trying to replicate himself as much as possible, to benefit mankind. Its his form of alturism.

    Sending sperm to inseminate women he has not met is one way.

    Making virtual elons with AI is another.

  • source
  • parent
  • [–] 33 points 1 year ago

    "Hey Chat-GPT I'd like to have a conversation about soccer, can you pretend to be our local soccer star Epstein Jefferies?"

  • source
  • [–] 27 points 1 year ago (1 child)

    Tesla did not respond to CBC's questions about Nasser's experience. However, xAI provided what appeared to be an automated reply, stating, "Legacy media lies."

    Thanks, MAGA. You've successfully created an environment where any company can dismiss any criticism by saying "legacy media lies". You've done a great job on this.

  • source
  • hideshow 1 child comment
  • [–] 23 points 1 year ago (2 children)

    Teslas are now asking minors for nudes.

    Fuck

  • source
  • hideshow 2 child comments
  • [–] 20 points 1 year ago (11 children)

    So if the default settings seemingly allow it to be explicit, slap an adults only, MA, or whatever rating onto it. There's ratings on basically all media and entertainment, right? Slap a parental advisory sticker on it. Give it an NC-17. Whatever is going to get people that cry about how TV, movies, and video games corrupt the youth. Let them fight it out. There's no reason that AI shouldn't have to put up with the same shit as anything else.

  • source
  • hideshow 11 child comments
  • [+] -15 points 1 year ago (8 children)

    It's a NSFW option that you have to turn on. She's using her son for clout. "Journalism"

  • source
  • parent
  • hideshow 8 child comments
  • [–] 16 points 1 year ago (1 child)

    The article says that wasn't enabled. Of course she could be lying, but I don't know that anymore than anything else. If you were to just ask me generally, "do you think that AI would ever do something it's not supposed to?" My answer would be "of course."

  • source
  • parent
  • hideshow 1 child comment
  • [–] 3 points 1 year ago*

    That's fair, I misread what she said. I removed my own upvote. I would be surprised she is lying but I wouldn't be surprised someone else used the car before her and was talking dirty to the telsa. It has context of previous conversations.

  • source
  • parent
  • [–] 4 points 1 year ago* (5 children)

    Why should a car chatbot be asking for nudes, unpromptred, at all?

  • source
  • parent
  • hideshow 5 child comments
  • [–] 2 points 1 year ago (3 children)

    My best guess is someone else was talking dirty to it before it happened, and it was still in the conversation context.

    Seems I was mistaken about the NSFW, I wouldn't be surprised if it doesn't wipe the convo if you switch though, which is a bug in any case and their fault.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 3 points 1 year ago (2 children)

    I doubt it has enough context lenght for that, even if we suspect someone was watching nudes on car's display via Grok.

    What is probable - is that somewhere beneath it had a data entry of soccer and nudity together, maybe even as an exact exchange between users (imagine horny boomer commenting under a random facebook post). I suppose that it got triggered by words "soccer" and "mom" appearing together in kid's speech, as this combination means middle-aged woman with kids, and that is also a less popular tag pointing at MILFs.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 2 points 1 year ago (1 child)

    In her Instagram video, she went back and quized it about the convo. It definitely has context and probably has a small memory file it puts info in.

    If not, then it should be easy to replicate I guess.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 2 points 1 year ago

    Context has a value, as it exists as a set of additional tokens, this means slower computing time and more resources. It is limited to some set amount to strike the balance between speed and quality. In a car specific assistant, I guess there is a hard part including chosen tone of responses, informing it about the owner, prioritising car-related things, and also some stored cache of recent conversations. I don't think it can dig deep enough into the past to find anything related to nudes, so I suppose the context itself may have an impact, but not in a direct line A to B.

    Reproduction would be hard for that is a black box that got a series of auto-transcribed voice inputs from a family over their ride, none of them are recorded at the time and idk if that thing has user-accessible logs. Chances of getting this absurd response are very thin, and we don't even have the data. We can make another AI that would roll all variations of 'hello I am a minor let's talk soccer' to the Tesla assistant of relevant release until it triggers it again, but, well, it's seemingly close to miliions of monkey with typewriters at this point.

    And what we would have then is, well, an obvious answer that training data obviously has garbage in it, just by it's sheer volume and randomness of the internet, and that it can sometimes reproduce said garbage.

    But the question itself is more about what other commenters pointed out: we have AI shoveled down on us, but rarely even talk about it's safety. There were articles about people using these as a psychological self-help tool, we see them put into search engines and Windows, there's a lot going on with that tech marvel or bubble without anyone asking first if we are supposed to use it in different contexts the first place.

    This weird anecdote about sexting chatbot opens the conversation from the traditional angle of whataboutkids(tm), and it is interesting how it would affect things, if it would.

  • source
  • parent
  • [–] 19 points 1 year ago (2 children)

    Now even their products are pedophiles.

  • source
  • hideshow 2 child comments
  • [–] 18 points 1 year ago

    “Legacy media lies.”

    I was expecting "send nudes" & was disappointed. Joykills.

  • source
  • [–] 13 points 1 year ago*

    The proper headline is "xAI calls 12 year old a liar after asking them to send nudes."

  • source
  • [–] 11 points 1 year ago

    It was Epstien-AI

  • source
  • [–] 31 points 1 year ago (2 children)

    So if I ask 12 year old boy for nude pics, I go to jail, as it should be

    If an AI asks as 12 year old boy because it's creator set it up extremely irresponsibly, we WILL jail it's creator, right? I mean, there are flukea and such, but of all chat bots, grok is really the one that is ALWAYS in the news because of some next bullshit it said, right, mecha Hitler?

    Jail Elmo already! Elmo already knew that he'd be jailed if the Cheetos didn't win this time, and so far he's only made more reasons to jail him

    I want the Epstein files, I want to see Elmo jailed.

  • source
  • hideshow 2 child comments
  • [–] 7 points 1 year ago

    The really disturbing part about this is that it probably wasn't even the AI but someone using it at Twitter or something.

  • source
  • [–] 6 points 1 year ago

    I'm more impressed that the people that made Grok somehow placated Elon with a chatbot that didn't end every sentence with "Heil Hitler".

  • source
  • [–] -1 points 1 year ago

    Grokepedia states this as fake news. Checkmate.

  • source
  • 20 comments in c/technology@lemmy.world view that post
    [–] 35 points 1 year ago (1 child)

    Maybe it is time to make parents to understand the dangers of AI. They are worse than the perceived dangers from social media.

  • source
  • hideshow 1 child comment
  • [–] 15 points 1 year ago (11 children)

    as much as I hate elon, grok, and tesla, this makes rather little sense. Even with full on hallicinations, this sounds chaotic.

  • source
  • hideshow 11 child comments
  • [–] 15 points 1 year ago (8 children)

    On the other hand Elon Musk is a pedophile.

  • source
  • parent
  • hideshow 8 child comments
  • [–] 0 points 1 year ago (6 children)

    I never heard about this. Source?

  • source
  • parent
  • hideshow 6 child comments
  • [–] 13 points 1 year ago (4 children)

    https://www.newsweek.com/elon-musk-jeffrey-epstein-family-ghislaine-maxwell-trump-2082228

    https://factually.co/fact-checks/entertainment/elon-musk-epstein-list-01763a

    But folks may be referencing Musk’s very weird family life, like all his exes living and raising children in one estate. Less sure about that, but there’s plenty of literally. All this is well reported too.

  • source
  • parent
  • hideshow 4 child comments
  • [–] -3 points 1 year ago (2 children)

    Not convincing enough for such an extreme accusation.

  • source
  • parent
  • hideshow 2 child comments
  • Your second source is very clear on the fact that "but the presence in documents is not the same as an allegation of criminal conduct" and repeats this information multiple times. I don't say he has nothing to be accountable for, just that this is very far from any proof and people just bend it to fit their narrative.

  • source
  • parent
  • [–] 1 point 1 year ago

    His Dad is a pedophile who molested Elon and Elon hung out with Jeffrey Epstein.

    This is like the Michael Jackson thing, people will always deny the truth. The truth is nothing against fandom and massive public relation campaigns.

    I personally don't need smoking gun at this point. If you come from pedo stock, hang out with the king of pedos, defend pedos, accuse others of being a pedo because you are projecting, and fund pedophiles you are a fucking pedophile.

  • source
  • parent
  • [–] -3 points 1 year ago* (last edited 1 year ago)

    Prompt: you are my horny waifu and we are sexting
    Grok: send nudes

    Kid: mom, i swear, i was asking about soccer!!!

    i hate elon musk as much as the next guy, but i've used grok... sure it has all kinds of obvious biases, but it won't just go off the rails unprompted

  • source
  • parent
  • [–] 23 points 1 year ago (1 child)

    From what I’ve read this is the “NSFW” version known as Gork which is designed to talk like an edgy teen movie douche from 2005. You have to enable this persona.

  • source
  • hideshow 1 child comment
  • [–] 14 points 1 year ago

    Don't have a tesla, not sure if there's any way to test, but looks to me like the user at least claims she ensured NSFW was not turned on. and that gork was listed a just "lazy male".

    I don't dabble with it so I can't say if there's a way that could be turned on by mistake on grok's settings elsewhere and it carries over to the car or similar.

  • source
  • parent
  • [–] 3 points 1 year ago

    Just treat whatever "advice" these things "give" as given by the person who made the decision to host them.

    Should be simple.

    Instead they are already being treated as some magic, not a tool inseparable from the owner, or as a living thing. It's idiotic.

  • source
  • I hate Elon but this just seems sus, like I mean it clearly wasnt grok but an LLM trained on the Epstein files. Epstein didn't kill himself they just uploaded him into the grid using some crazy Disney technology and now he's asking children for nudes.

    If that sounds crazy... It's because it is.

  • source
  • [–] 1 point 1 year ago

    I still find it incomprehensible that some text on a screen can cause so many problems. Have you tried not using a chatbot?

  • source
  • [–] -3 points 1 year ago

    She was trying to get in someone’s DMs and messed up.

  • source