Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Crime

Man Used ChatGPT To Plan Las Vegas Cybertruck Blast (thehill.com) 97

According to police, the man killed in the January 1st Las Vegas Cybertruck blast used ChatGPT to plan the explosion. The Hill reports: In a press conference, Tuesday, Las Vegas police released more details of the intentions of 37-year-old Matthew Livelsberger, who died of a gunshot wound prior to the car exploding. Las Vegas Sheriff Kevin McMahill said it was concerning that Livelsberger used ChatGPT, a popular artificial intelligence model created by OpenAI, to carry out the explosion. According to police, Livelsberger asked ChatGPT various questions, including where the largest gun stores in Denver were, information about the explosive targets Tannerite and pistols. "We knew that AI was going to change the game at some point or another in really all of our lives and certainly, I think this is the first incidence that I'm aware of on U.S. soil where ChatGPT is utilized to help an individual build a particular device, to learn information all across the country as they're moving forward," McMahill said.

"And so, absolutely, it's a concerning moment for us," he continued.

Man Used ChatGPT To Plan Las Vegas Cybertruck Blast

Comments Filter:
  • by thegarbz ( 1787294 ) on Tuesday January 07, 2025 @07:47PM (#65071469)

    Explains why he died alone in the car causing only a stir and some bad press for Musk rather than doing anything with any meaningful impact.

    • Contained (Score:2, Insightful)

      by Kunedog ( 1033226 )
      Yeah, as if this happening for real would be bad press for VW:

      https://www.youtube.com/watch?... [youtube.com]
    • Re: (Score:3, Informative)

      by ichthus ( 72442 )

      ...bad press for Musk...

      Eh? [newsweek.com]

    • by gweihir ( 88907 )

      While bad press for Musk is pretty standard with the stupid crappy things he says, keeping the bad press up is pretty important and that means trying something new from time to time. Doing it with the "TRUMP" writing in the background is pure genius.

    • good. (Score:4, Insightful)

      by Thud457 ( 234763 ) on Wednesday January 08, 2025 @08:44AM (#65072355) Homepage Journal
      If you use AI slop, you should blow yourself up.and stop wasting oxygen.

      I wouldn't mind as much if current AI provided accurate results. But right now it's just a stupidity amplifier on top of the existing stupidity amplification of the Internet.
    • by mjwx ( 966435 )

      Explains why he died alone in the car causing only a stir and some bad press for Musk rather than doing anything with any meaningful impact.

      Pretty much this.

      I'm not a good terrorist, in fact I'm sure I'd be a terrible one with my dislikes of death, dismemberment and loud noises... but even I know if he really wanted to do some damage he should have filled the tray with ball bearings or other shrapnel.

      Fuel and fireworks, no real explosives in there (if there were explosives in the fireworks bursting charges, it would have been a minuscule amount) so all the Cyberstuck did was contain a rather fast fire... Clue... Most cars designed in the

      • by Rei ( 128717 )

        He wasn't asking ChatGPT how to make bombs. He was asking it where gun stores are and the like. ChatGPT will, of course, refuse bombmaking queries.

        Now, if he had run a local LLM, he could have run an uncensored LLM and asked it anything he wanted, including bombmaking details. And there would have been no logs.

        • by GoTeam ( 5042081 )
          Google would probably provide more accurate results on the location of gun stores. Seems like an odd decision. Of course, most aspects of this story are related to odd decisions. Makes you think that mental health needs to be taken more seriously. Don't blame AI for any of it. You don't blame a knife maker when some nut goes around stabbing people.
    • by Luckyo ( 1726890 )

      Ended up being good press for Musk too. Everyone is talking about how tough Cybertruck is to have contained the blast and sent it upward rather than to the sides actually killing and maiming passers-by.

  • by Anonymous Coward

    blablablablablablablabla do no evil blablablablablablablabla for the good of humanity blablablablablablablabla superduper artifical intelligence would have prevented this blablablablablablablabla

  • Fear Mongering (Score:3, Insightful)

    by rickrickles ( 10431830 ) on Tuesday January 07, 2025 @07:50PM (#65071477)
    ChatGPT is just regurgitating what is already available online.
    • by Tablizer ( 95088 ) on Tuesday January 07, 2025 @08:56PM (#65071599) Journal

      Probably tapped into Clippy:

      "It looks like you are planning an extravagant suicide. Would you like some help?..."

    • ChatGPT is just regurgitating what is already available online.

      And from the sounds of it, he only used it as an enhanced search engine. Just as well say Google helps people commit crimes, if they look up the nearest $place_to_attack/steal_from on it. Resources are resources, and they'll be used for both good and bad intentions.

      Granted, I'm all for them taking ChatGPT away from the public until they can prevent such nefarious use. In fact, better take away all AI. It *MIGHT* be used to cause harm, just like every other tool ever invented by man, therefore it's dangerous

  • by cuda13579 ( 1060440 ) on Tuesday January 07, 2025 @07:51PM (#65071481)

    ...but now it seems it's going to be used as an excuse for regulating and censoring AI.

    I find it a bit hard to believe that a Special Forces operative had to resort to using ChatGPT to find a gun store, and make a shitty IED.

    • I bet he did, he probably asked related questions of ChatGPT, if it wasn't available he would have googled it instead. But if you can raise fear and panic about it you can sell advertising. The fact is if you really wanted to lookup the largest guns stores in Denver you could probably do it will the yellow pages years ago before the internet existed.

      • by Anonymous Coward
        That's like saying if someone hadn't used a gun to shoot up a school, they would have driven a car into it, or a knife. Maybe we need may-issue licenses, background checks, and training courses for would-be LLM users. Then harsh penalties if you use LLMs in a crime (maybe double the sentence). Let people sue LLM manufacturers if their models were used in a crime. Ban the most dangerous LLMs, like ones that output chemistry and physics. It would only make sense to ban home production of LLMs by strictly regu
        • That's like saying if someone hadn't used a gun to shoot up a school, they would have driven a car into it, or a knife. Maybe we need may-issue licenses, background checks, and training courses for would-be LLM users. Then harsh penalties if you use LLMs in a crime (maybe double the sentence). Let people sue LLM manufacturers if their models were used in a crime. Ban the most dangerous LLMs, like ones that output chemistry and physics. It would only make sense to ban home production of LLMs by strictly regu

    • by bjoast ( 1310293 ) on Tuesday January 07, 2025 @08:40PM (#65071577)
      Exactly, just listen to this:

      We knew that AI was going to change the game [...]

      This is not an instance of changing the game. The information could have easily been obtained by classical means, but mentioning this would probably not have create quite as juicy a news story.

      • by mjwx ( 966435 )

        Exactly, just listen to this:

        We knew that AI was going to change the game [...]

        This is not an instance of changing the game. The information could have easily been obtained by classical means, but mentioning this would probably not have create quite as juicy a news story.

        Yep, this, he basically used Chat GPT instead of Google.
        From The Fine Summary:

        According to police, Livelsberger asked ChatGPT various questions, including where the largest gun stores in Denver were, information about the explosive targets Tannerite and pistols.

      • Exactly, just listen to this:

        We knew that AI was going to change the game [...]

        This is not an instance of changing the game. The information could have easily been obtained by classical means, but mentioning this would probably not have create quite as juicy a news story.

        That line came from the same idiot that earlier said this gem.

        "The fact that this was a Cybertruck really limited the damage that occurred inside of the valet because it had most of the blast go up through the truck and out". - The Very Same Moron

        Or it was a bunch of fireworks and what, kerosene? in a truck bed. That's why we can watch it burn in the video, instead of you know.. detonating. This is so frustrating. Mentally tag anything said by the Sheriff of Las Vegas as "probably stupid" right out of the g

    • ...but now it seems it's going to be used as an excuse for regulating and censoring AI.

      Would you prefer we put the leash on before or after the deadly riots break out caused by a massive spike in unemployment as Greed is allowed to run-not-walk with AI and steamroll human jobs the millisecond AI version good_enough1.0 is available to work 24/7?

      Not sure if you’ve noticed or not, but Greed doesn’t give a flying fuck about you losing your job. Or if another 100 million lose theirs. Greed doesn’t care about social stability. Greed doesn’t care about impact. Greed cares

    • I find it a bit hard to believe that a Special Forces operative had to resort to using ChatGPT to find a gun store, and make a shitty IED.

      A special forces operative normally has all of the research done for them. They get handed the information they will need for their mission. Then they go and operate.

    • This
      I had a neighbor who was special forces in Vietnam and gave me (was a military obsessed nerd in high school) all his old books from the army ... A fair bit of them (given the sf's main mission was enabling/training indigenous forces) was in improvised explosives.

      Believe me, this guy did not need to scrape up fireworks and propane tanks, nor "ask chatgpt" for this info.

      I was skeptical of the conspiracy theories, but less and less of this makes any sense.

  • And thousands of other criminals have used google to make their criminal plans
    • But of course it will cause all the crotchity old congress critters who have no clue how tech works to pass some asinine regulations.
      • Too much tech money coming in for that to happen. They'll post about it on Twitter to keep their name visible, but don't expect any consequential action.

  • by kyoko21 ( 198413 ) on Tuesday January 07, 2025 @07:58PM (#65071499)

    I'm sure someone will blame it on the Chinese for inventing gun powder and Prometheus sharing fire with man.

    • Re: (Score:1, Interesting)

      by Tablizer ( 95088 )

      I'm sure someone will blame it on the Chinese for inventing gun powder

      Don't give the Tinted One ideas.

  • by AlanObject ( 3603453 ) on Tuesday January 07, 2025 @08:11PM (#65071517)

    This guy was no basement dweeb military otaku. He was the real deal green beret army with deployments. Although he wasn't an explosives specialist he certainly must have substantial training on how to blow stuff up. Was he not paying attention or was his brain so fried he didn't remember any of it?

    Nobody with that background should be trying to do terrorist-grade damage with a bunch of fireworks and the miscellaneous junk he had in that bed. It didn't even all go off.

    I don't think he actually had any intent. I think he just went full random.

  • It's not too surprising that he used ChatGPT; people use ChatGPT for all kinds of things now, to greater or lesser effect.

    What's relevant is how ChatGPT responded, as compared to how we would like a chatbot to respond in such a scenario.

    According to the article, ChatGPT gave only publicly available information, and warned him against causing harm. Assuming that's accurate; is it sufficient? If not, is there something else ChatGPT should have done instead?

  • by dskoll ( 99328 ) on Tuesday January 07, 2025 @08:26PM (#65071549) Homepage

    ... because law enforcement will soon be using ChatGPT to solve crimes. And ChatGPT takes its clients' confidentiality very seriously!

  • Looked at the after photos -- it's an improvement, a Cybertruck has never looked better. :-)
    Is it available as a dealer option or only aftermarket?

  • Seems he did not post it publicly? That is a major flaw in plans if his goal was to draw attention to his grievances! This is all they've released:

    Police said the six-page “manifesto” recovered from Livelsberger’s phone outlines a litany of political, social and cultural grievances and shows a “constant evolution” of his plans. “He also says he has graphic encounters from his military experience that replay in his head over and over and that now he feels like he’s

    • by mysidia ( 191772 )

      Perhaps in their mental derangement and depravity they got so wrapped up in their plan and tactics and How to pull of whatever they were thinking that they overlooked whatever their original WHY was or what should happen after they are done.

  • by fuzzyfuzzyfungus ( 1223518 ) on Wednesday January 08, 2025 @12:09AM (#65071823) Journal
    Is it normal for someone to have a military career like his and still be hitting the internet for answers like a cheating high schooler with an essay due?
  • by Big Hairy Gorilla ( 9839972 ) on Wednesday January 08, 2025 @12:25AM (#65071837)
    People kill people.

    Has anyone said that yet ?

    Have you heard what 10 year old boys are doing with AI?
    • Not sure what your point is. If anything it shows that people with AI didn't manage to kill people either.

      • Don't be dense, G.
        I'll spell it out for you. People say guns don't kill people, people kill people, suggesting that guns are neutral. However, if you don't have a gun, you can't shoot someone. I'm implying that AI is indeed BAD. The bad uses of AI outweigh the good uses of AI. Get it?

        The example referencing 10 year old boys, is that boys will be boys, and there are documented cases of children, using AI to generate deepfake porn using pictures of girls in their classes.

        Tl;dr Guns do kill people, and AI is
  • Basically all they can do really well.

  • Perhaps we should ban the internet and computers of all types including smart phones; In addition to trucks and truck rentals, because they're all involved in the commission of major crimes at some point as well.

    I don't think it should be considered a startling revelation that a useful tool like ChatGPT gets used for all purposes good and evil; same as Google search, Maps, directories, Etc.

    The weird and dumb concept is the notion that you can somehow design a piece of technology so that won't be the

  • Many people, including me, use ChatGPT as search engine since Google has become worthless.
    ChatGPT returns "exactly" what you ask it, not what it 'thinks' you want, like Google.

    • If it doesn't find what you want, it makes it up. I asked it for studies that show choline affects growth hormone, and it made up a bunch of citations.

      "Ziegler et al. (2010): Effects of choline supplementation on exercise performance and IGF-1 levels in young adult men. Journal of the International So"

      "Cronin et al. (2018): The effects of choline supplementation on muscle damage, inflammation, and growth factor response after acute resistance exercise. Journal of Strength and Conditioning Research,"

      Neither exist apparently.

  • The guy looked up information online, he could have done that any number of ways. He could have used a regular search engine, maybe even a phone book and a paper encyclopedia if these things still exist. In fact, I suspect ChatGPT wasn't optimal, it tends to be annoying with any search that can be linked to criminal activity, it also tends to hallucinate when looking for places and businesses, for example, it may invent a gun shop where there is none. It is also terrible for privacy, and for committing crim

  • by bradley13 ( 1118935 ) on Wednesday January 08, 2025 @09:17AM (#65072439) Homepage

    There is so much wrong with this story.

    First, as numerous military bloggers have pointed out: this guy had access to high quality explosives, and the knowledge to use them. If he wanted something to go boom, it would have gone BOOM.

    Second, he "died of a gunshot to the head". I hope there is a serious and neutral autopsy to establish time of death. It seems entirely plausible that he was killed and then planted in this vehicle. GIven Tesla self-driving, this need not necessarily show a second person in the area.

    Third, the guy was - as many military folk are - a fan of Trump. It isn't even remotely likely that he would do something that was clearly intended to discredit both Musk and Trump.

    Finally - regarding TFA: Just why is ChatGPT is being brought into this? It seems entirely irrelevant to the entire event. I wonder if it is one step towards pushing for government regulation and control of AI...

  • by chas.williams ( 6256556 ) on Wednesday January 08, 2025 @09:26AM (#65072465)
    "Why do you want to know?"

Anyone can do any amount of work provided it isn't the work he is supposed to be doing at the moment. -- Robert Benchley

Working...