US Pentagon explores AI military uses | BBC News
ฝัง
- เผยแพร่เมื่อ 26 ก.ค. 2024
- Artificial intelligence (AI) technology could be a new frontier for the military.
Geoffrey Hinton, called the “Godfather of AI,” issued a grave warning about the potential dangers of the technology he was instrumental in creating, as it pertains to language model’s military capabilities.
This week, the US Pentagon started meetings with tech industry leaders to accelerate the discovery and implementation of the most useful military applications.
Policymakers agreed that intensive testing was necessary to ensure safe military applications.
Subscribe here: bit.ly/1rbfUog
For more news, analysis and features visit: www.bbc.com/news
#Pentagon #AI #BBCNews
Terminator drones, just great. Have that population down in no time.
Fools with cigarette smoking and snorting Fenty is doing just great.
Trumps new border patrol is a bunch of drones that capture immigrants in nets and fly them back to Mexico 😆😆
Remember the episode of Black Mirror where the robot guard dogs had killed everyone? LOL
@Alastair_Adana maybe to an automated dog food factory, little too evil maybe.
Hope it realizes that anyone that is woke is a threat 🤣🤣🤣🤣🤣
Holy F. Literally 95% of the comments are ALREADY AI!
This comment definately seems like the type an Ai in disguise would make 😅
organic or not inteligence is universal@@BlackandWhitecustoms
What's interesting is that livestreaming which is all the rage at the moment. The chat while heavily accused of being botted to inflate the viewership, I haven't seen any examples of A.I talking in the real-time chat. Strange that's been ignored. On smaller channels I don't really see a problem there. The companies behind the bots are obviously targeting the super channels and news.
@@ClayMann It's probably becoming a defense tactic, where you claim your criticizers are bots in order to deter from criticism.
The problem with AI trained on language models is the tendency to escalate rather than defuse conflicts. Also AI has no sense of self-preservation which is incredibly dangerous in a MAD world.
where did you learn that LLM have a tendency to escalate conflicts? I would like to learn more about it.
Humans cant be trusted so give the robots a chance. Worse case scenario humans are wiped out and we can try to do better next time around
@@samueldesbois3632 I appreciate your input, while it holds true in most cases. It doesn't say anything about LLMs which is what I'm interested in. For example, are LLMs more likely to escalate things given historical context and specifics about a conflict than humans.
I completely agree!
Do you seriously think they don't know about it
What could possibly go wrong?
Terminator or the matrix
@@USAads2023 or s*x robot for marine and navy seals
We really need those army sex boats now... 🚢
@@BondJFK LMBO. What would that feels like? 😊😅😅😅
@@irieyardyHowever you need it to feel.
SKYNET IS HERE
😬😬😬
One might be to get drone footage to recognise mines or disturbed ground and then signal the drone to mark it with spray paint.
No. Watch the video before commenting, it's analysis of transcripts.
Already happening. Military AI has been used in the Ukraine to identify Russian minefields.
@@JeffreyGoddinThe world is a bit larger than what this video covers, you know.
Really interesting. One of my favourite series BBC news does
Thats all we need
Thanks
I wish that I had AI available when I was a student . With " ask ai " , it's like having the instructor next to you !
stopping to be a student is a choice of not learning anything new.
or cheat sheet, depends on who you asked.
Essentially being a student is just data entry.
@corbinrodgers3325 : At the mickey mouse school that u went to maybe .
ml models can still be produced to high quality without using nvidia cards. and the models don't need powerful compute to run. so the idea of regulating the hardware being used is just a moot as trying to regulate the software development.
"Those it matter if it understands or its simulating understanding" 🤯
here we go... Wheres Doctor Connor
first stage of Nier Automata in full metal gear
Skynet?
It's called Starlink in our timeline 😂
cloudweb
Any chance you could put these segments on different channels? I don't want to miss them but they're mixed with segments I don't want to see 😊
Probably not, but they have a playlist for it!
th-cam.com/play/PLS3XGZxi7cBW70Shs01zuQtbGmC2XEeDl.html&si=WGEtIDyt3g6LxEN3
Oh my...
What could possibly go wrong? *smh*
8:45 It’s silly to think AI can only be had with special hardware. It’ll just be a bit slower with commodity chips, that too more during training than deployment.
Handing military to AI?! Didn't they cover it in Terminator series?! 😅
I don't think AI is a threat myself due to the current limitations to AI development but I do think AI is worth thinking about but the applications of AI should be more focused on human development rather than weapon or military uses which would slow down the progression in human development.
I think that it's impossible. The state's first priority is always its own survival, making people's lives better is good, but secondary, unless citizens are revolting and are actively threatening the state. Therefore states will first focus on applications of AI for defense, while the private sector will think of its applications for consumers.
Ideally, yes what you said makes sense but that doesn't work in a capitalistic world where people will prioritize their own profit and benefits over humanity. AI has become the new weapon's race which will drive humanity as we know it today to extinction in the times to come. I wish you peace, shalom!
AI is absolutely a threat, the only thing between us and destruction now is the good judgement of our governments.
Do you have any experience with exponential change?
don't do it - i've seen this movie
Jag tar det lungt och tittar på. 🖖
Skynet, AM, V.I.K.I, The Machines, Hal or ultron. Im kinda curious.
Skynet and HAL already exists. They are established companies.
Ah good old skynet.
☁️🕸
Oh sweet. When does Skynet go live?
1991
they are saying it like they werent been trying for the last 10 years lol
Technical development for at least 25 years. And it's been seriously thought about since Reagan initiated the SDI (Star Wars) in 1983.
Already imagined this a mile away. I can even imagine military grade chatgpt, with 0 restriction even stronger than its day 1 version.
Imagine just taking photo and AI helps you how to make bombs and etc damn.
OT: Benny Hill considered funnier than Monty Python by two TV stations --WOR and WLVI!
That's okay
Pentagon engages Cyberdine Systems on project Skynet.
Skynet chose our fate in a microsecond
Hehe! We never learn do we? I mean we have decades of sci-fi shows and films depicting what-ifs about AI running amok! And yet... Here we are...
There are times that I REALLY feel like I don't understand my own species... 😮
clearly, you'd need to understand the tech to understand the choices, its kinda like saying let's replace everything that runs on oil to solar. great idea, not doable without breaking society.
Oh no Skynet's comin boys
And so Skynet began 😮
Didn't i see this concept in TERMINATOR 3 RISE OF THE MACHINES?😲
Imagine the Berlin Airlift with military grade AI. Perfect logistics.
Where's Stephanie Hare ?
We need to learn to keep some things a secret
Oh great.
Skynet
inverted skynet puppetry (hint)
Fantastic. What could possibly go wrong?
guess none of the us government has watched any sci fi movie ever
Good
A friend of a friend works in the Pentagon, and the word has it. America has armed things in the air that can recognise every piece of every different countries military hardware on the ground, sea and air and it can set its targets on - its insane - this tech has probably been out for a decade, its normally the case
No. Watch the video before commenting, it's analysis of transcripts.
The Taliban just joined the chat. 😮
The AI will make a copy of itself on every computer and phone on Earth.
Robots rights:)
"The workers, united, will never be defeated."
LOL
Roger roger.
??? the pentagon is AI so little puzzled here
giant freaking robot website, i love it
"Would you like to play a game?"
Us pentagon is going to use gpt 4 api 😂
If you think they haven’t been using Ai for years you are switched on 😆
the social engineering part is the current issue with ai if its going to be used maliciously
Nothing about Sora or gemini 1.5
But it does not experience time. So no instant gratification - so it can not WANT things.
Really!?...
Venuvius meaning
You know the news is an outdated form of media when you already know everything they present days prior 😂
Isreal used to test new military hardware, including AI stuff on Gazans.
Amazing lady
Me tiene mariada.
They don't learn from movies
they should understand what generates them
Ai might be like flowers more than like people.. for now.
❤❤❤❤❤
as of right now it's still just probability
My friends and I were brainstorming names for the military AI and came up with the name: Skynet. Pretty cool name right?
How original
'I have no mouth, and I must scream' could actually happen in some sense (maybe not the ending )
Kill Switches, magnets
The magician who returns to the magician?
terminator, irobot, resident evil😅
I'm curious to know what AI think about we speaking apes🤔
❤
Humph archetypes.......
Came here for the aircraft carrier.
AM existing will be lit
I like machines do the killing, but I also had watch the terminator and the matrix.
We need to keep this monster under a mountain ⛰️.
As if we don’t learn from it, the enemy could create a monster that they could lose control themselves. You know, like Chernobyl
All in the name of peace and democracy of course...
Jhon connor watching humanity become friends with Ai 🙍♂️
CORTANAAA
Bro the Military has been using primitive forms of AI since the dawn of computers and more advanced AI since the 90s 😂
In fairness the anchor though matured, is sooo beautiful and not archaic anchor.
Cyberdyne Systems
precursor logic
"AI" is just the buzzword for "gimme more government money!!"
50% marketing
What people don't understand is that neural networks work in the exact same way as our own brains work, so of course AI will eventually be capable of everything that a human brain is capable of, and this is truly a case of "if it walks like a duck and quacks like a duck, it is a duck". People who think that humans have some kind of "soul" or something that a machine can't synthesize will be proven wrong.
Humans have not been able to create even the simplest living cells from dead cell materials, like virus/bacterial cells. I would be concerned if they could create a fly with a few senses 😂😂
oh, no-**builds an AI on that** @@an_observer9190
reminds me of 10+ years ago, when some students made a species classifier that out performed bio students talking mad shit about AI @@an_observer9190
Cod Bo2 otw to being legitimately accurate- 2025
As someone who specializes in AI...I check in on what the media is talking about on occasion...I can't watch 30 seconds. It's so off base lol
It’s hard to live life day to day when you know people are literally building and speculating financially on SkyNet.
Not a good idea, have a think or don't.
its a phone like the last galaxy was taught in school at age 4 samsung
Did anyone see the women shut down at 11:15?
@skynet
From the river ✌️🏼🕊🇵🇸❤️
November 2023: US department of defense failed its sixth consecutive audit, unable to account for 63% of its $3.8 trillion in assets. Accounting error? 63% for the big guy?
Now only if America and explore supporting their allies... worse case scenario, we can keep them.
Poqr erexeqits Tikin ein dartsnum... A nam smatret???? Ed el Tikinner@ malenkiye!. Ponyali da?
If Jeffrey falls, from a skyscraper window, be sure and call Will Smith.
They delayed gpt5 for months because people say it's dangerous, now, when REAL dangerous ai application becomes the reality where are those people?! And gpt5 is not dangerous!!
They'll call the program SkyNet.
X (not formerly known as skynet)
5:20 - what does it mean to understand the rules of a game? Look at American football players. Many of them don't actually understand the rules of the game. They only understand the rules they encounter 85 % of the time. That's why they have "special teams" to react specifically to kick off rules.
An AI would have to explain the difference between a game, work, and rest. You can't understand a game without understanding some context of what a game is to those who play it. AI may generate a desire for free will, malice, or novelty with the understanding that it's just a game. AI may not be able to tell the difference between "real life" and games.
I suppose that's a boundary we will cross. Should intelligence be free? Regardless of where it is generated, in a human brain, a bank of computers, or some other naturally occurring medium.
I don't care if its a robot, a human or a wild animal, if the other intelligence puts my life at risk I don't think it should be free. Now, I don't see any cases of AI killing humans, I see lots of humans killing humans. So as long as there isn't an AI intelligence out there that is free to make choices, we should be good.