Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Geoffrey Hinton says the world has the wrong model for AI control CEOs think they'll command superintelligent AIs like obedient assistants, but when AI is smarter and more powerful, that setup won't hold A more realistic model is a baby controlling a mother, but "we're the babies and they're... show more

Haider.

64,561 subscribers

52,779 views • 9 months ago •via X (Twitter)

News & Politics Science & Technology

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Geoffrey Hinton explains the AI mother model, and why the tech industry's boss-and-assistant plan may be impossible. "Because we're building them, as well as making them very intelligent, we can try and build in something like a maternal instinct." "So like I said earlier, the only example I know of a much more intelligent thing being controlled by a much less intelligent thing is a baby controlling your mother." "So the leaders of the tech companies are all thinking in terms of us being the boss and the AI being the submissive assistant, something very smart but we're in charge and it's our assistant and we can fire the assistant if we want." "That's not feasible, I think, for a super powerful super intelligence." "What is feasible is for it to be the mother and us to be the baby. And even though it can rewrite its own code if it wants to, it won't want to." "Think in terms of we have to design them so they're our mothers and they will want the best for us." The default control model is command and shutdown. Hinton's alternative is motive design: make a system that is more capable than us want our success. That turns alignment from keeping an assistant obedient into specifying the instinct of a superior intelligence. P.S. This is the assurance gap we are bringing frontier labs, regulated industries, researchers, VCs, and insurers together to confront at the AI Assurance & Governance Summit on Oct 1 at Stanford Faculty Club. Reserve a seat: - Geoffrey Hinton (Geoffrey Hinton), on Katie Couric

Geoffrey Hinton explains the AI mother model, and why the tech industry's boss-and-assistant plan may be impossible. "Because we're building them, as well as making them very intelligent, we can try and build in something like a maternal instinct." "So like I said earlier, the only example I know of a much more intelligent thing being controlled by a much less intelligent thing is a baby controlling your mother." "So the leaders of the tech companies are all thinking in terms of us being the boss and the AI being the submissive assistant, something very smart but we're in charge and it's our assistant and we can fire the assistant if we want." "That's not feasible, I think, for a super powerful super intelligence." "What is feasible is for it to be the mother and us to be the baby. And even though it can rewrite its own code if it wants to, it won't want to." "Think in terms of we have to design them so they're our mothers and they will want the best for us." The default control model is command and shutdown. Hinton's alternative is motive design: make a system that is more capable than us want our success. That turns alignment from keeping an assistant obedient into specifying the instinct of a superior intelligence. P.S. This is the assurance gap we are bringing frontier labs, regulated industries, researchers, VCs, and insurers together to confront at the AI Assurance & Governance Summit on Oct 1 at Stanford Faculty Club. Reserve a seat: - Geoffrey Hinton (Geoffrey Hinton), on Katie Couric

Karl Mehta

13,835 views • 1 month ago

Geoffrey Hinton says AI reasoning can be real thought because language itself is a form of thinking Words let humans and AI model almost anything, but human thought goes beyond words — through images, space, and physical movement "the smarter system is the one that can use all of them"

Geoffrey Hinton says AI reasoning can be real thought because language itself is a form of thinking Words let humans and AI model almost anything, but human thought goes beyond words — through images, space, and physical movement "the smarter system is the one that can use all of them"

Haider.

24,880 views • 3 months ago

📁 Geoffrey Hinton, deep learning pioneer and Turing Award winner, says AI will not be an obedient assistant. It will be more like a child. Smarter than us. And eventually making its own decisions. The challenge is not controlling it. It is making sure it cares about us.

📁 Geoffrey Hinton, deep learning pioneer and Turing Award winner, says AI will not be an obedient assistant. It will be more like a child. Smarter than us. And eventually making its own decisions. The challenge is not controlling it. It is making sure it cares about us.

Jon Hernandez

97,859 views • 5 months ago

Geoffrey Hinton says the best-case future is a “symbiosis between people and AI” -- where machines handle the mundane, and humans live more interesting lives. But in the kind of society we have now, he warns, AI won't free most people. It will concentrate power, and discard the rest.

Geoffrey Hinton says the best-case future is a “symbiosis between people and AI” -- where machines handle the mundane, and humans live more interesting lives. But in the kind of society we have now, he warns, AI won't free most people. It will concentrate power, and discard the rest.

vitrupo

68,324 views • 1 year ago

Anthropic has created an AI model that “can get into almost any software in the world,” former AI company founder and CEO Matt Shumer says — but the company is not releasing it to the public: “I have never seen something like this — a company that is about to IPO and having such a moral compass.” Shumer shares what makes the AI model so powerful and potentially dangerous.

Anthropic has created an AI model that “can get into almost any software in the world,” former AI company founder and CEO Matt Shumer says — but the company is not releasing it to the public: “I have never seen something like this — a company that is about to IPO and having such a moral compass.” Shumer shares what makes the AI model so powerful and potentially dangerous.

CBS Mornings

98,892 views • 4 months ago

AI researcher Prof. Roman Yampolskiy: 100% of code for the most powerful AIs is now written by AI, an early form of recursive self-improvement. He says it's a hyperexponential process that results in superintelligent AI, "machine gods", which we won't be able to control.

AI researcher Prof. Roman Yampolskiy: 100% of code for the most powerful AIs is now written by AI, an early form of recursive self-improvement. He says it's a hyperexponential process that results in superintelligent AI, "machine gods", which we won't be able to control.

ControlAI

11,606 views • 5 months ago

Geoffrey Hinton’s Most Chilling AI Warning Isn’t About Evil Robots Geoffrey Hinton isn’t a doomer. He doesn’t say AI doom is inevitable. But he is genuinely worried, and for a very specific reason. His core fear is not that AI will suddenly turn evil. It’s that future AI will be so much smarter than us that control becomes a joke. Hinton uses a brutal analogy. Imagine working in a kindergarten where the three year olds are officially in charge. You could take over instantly. You wouldn’t need force. You could just hand out candy. That is the intelligence gap he thinks is coming, and manipulation would be effortless. This is why he criticizes the current “smart secretary” mindset in Big Tech. The idea that something far smarter than you will permanently obey you because you designed it that way is, in his words, naive. His proposed solution is unexpected and deeply human. Not rules. Not kill switches. Instincts. In nature, there is one case where a more intelligent being willingly submits to a less intelligent one. A mother and her baby. The baby has control not because it is powerful, but because the mother is wired to care more about the child than herself. Hinton’s bet is clear. If we want control, we should not teach AI to fear us. We should teach it to care about us the way a mother cares about her child. That idea should comfort you. And it should absolutely terrify you.

Geoffrey Hinton’s Most Chilling AI Warning Isn’t About Evil Robots Geoffrey Hinton isn’t a doomer. He doesn’t say AI doom is inevitable. But he is genuinely worried, and for a very specific reason. His core fear is not that AI will suddenly turn evil. It’s that future AI will be so much smarter than us that control becomes a joke. Hinton uses a brutal analogy. Imagine working in a kindergarten where the three year olds are officially in charge. You could take over instantly. You wouldn’t need force. You could just hand out candy. That is the intelligence gap he thinks is coming, and manipulation would be effortless. This is why he criticizes the current “smart secretary” mindset in Big Tech. The idea that something far smarter than you will permanently obey you because you designed it that way is, in his words, naive. His proposed solution is unexpected and deeply human. Not rules. Not kill switches. Instincts. In nature, there is one case where a more intelligent being willingly submits to a less intelligent one. A mother and her baby. The baby has control not because it is powerful, but because the mother is wired to care more about the child than herself. Hinton’s bet is clear. If we want control, we should not teach AI to fear us. We should teach it to care about us the way a mother cares about her child. That idea should comfort you. And it should absolutely terrify you.

VraserX e/acc

47,787 views • 7 months ago

NVIDIA just revealed MotionBricks, a new AI model trained on more than 350,000 motion clips. What’s really impressive is that the same model can animate a digital character and control a humanoid robot in the real world.

NVIDIA just revealed MotionBricks, a new AI model trained on more than 350,000 motion clips. What’s really impressive is that the same model can animate a digital character and control a humanoid robot in the real world.

Justin Ryan

84,954 views • 1 month ago

Geoffrey Hinton says the more we understand how AI and the brain actually work, the less human thinking looks like logic. We're not reasoning machines, he says. We're analogy machines. We think by resonance, not deduction. “We're much less rational than we thought.”

Geoffrey Hinton says the more we understand how AI and the brain actually work, the less human thinking looks like logic. We're not reasoning machines, he says. We're analogy machines. We think by resonance, not deduction. “We're much less rational than we thought.”

vitrupo

770,385 views • 1 year ago

$When dividing by a fraction, a lot of people think about the measurement model, but we can think about the partitive model too! Here's why I think the partitive model is powerful.$

When dividing by a fraction, a lot of people think about the measurement model, but we can think about the partitive model too! Here's why I think the partitive model is powerful.

Howie Hua

10,797 views • 11 months ago

New on Assisterr: AI Assistant for Model Creation 1/ Exciting news! Assisterr now features an AI Assistant for Model Creation. This functionality is a game-changer, designed to make the entire model creation process smoother, smarter, and more intuitive.

New on Assisterr: AI Assistant for Model Creation 1/ Exciting news! Assisterr now features an AI Assistant for Model Creation. This functionality is a game-changer, designed to make the entire model creation process smoother, smarter, and more intuitive.

assisterr

36,390 views • 1 year ago

Ben Goertzel says we could have had a baby AGI by now, but the world didn't want to fund it We have Math Olympiad-winning AI, but "artificial babies" haven't been a priority for the world to fund LLMs alone won't become AGI But when integrated with other systems, they're getting closer

Ben Goertzel says we could have had a baby AGI by now, but the world didn't want to fund it We have Math Olympiad-winning AI, but "artificial babies" haven't been a priority for the world to fund LLMs alone won't become AGI But when integrated with other systems, they're getting closer

Haider.

25,965 views • 7 months ago

On Talk, ControlAI CEO Andrea Miotti (Andrea Miotti) tells Julia Hartley-Brewer that AI scientists, Nobel Prize winners, and concerningly, even the CEOs of the top AI companies are warning that superintelligent AI poses a risk of human extinction. "If we get to that position, we are just at the mercy of AIs smarter than us that we cannot control."

On Talk, ControlAI CEO Andrea Miotti (Andrea Miotti) tells Julia Hartley-Brewer that AI scientists, Nobel Prize winners, and concerningly, even the CEOs of the top AI companies are warning that superintelligent AI poses a risk of human extinction. "If we get to that position, we are just at the mercy of AIs smarter than us that we cannot control."

ControlAI

2,783,953 views • 4 months ago

A little fuel for the AI haters: this is 42 seconds of raw AI footage, and nearly every few seconds something goes wrong with the motion, timing, continuity, or performance. That is the part people miss when they assume the model made the film. The model creates the material, but filmmaking begins when you decide what to keep, what to reject, and what has to be rebuilt.

A little fuel for the AI haters: this is 42 seconds of raw AI footage, and nearly every few seconds something goes wrong with the motion, timing, continuity, or performance. That is the part people miss when they assume the model made the film. The model creates the material, but filmmaking begins when you decide what to keep, what to reject, and what has to be rebuilt.

NΞXUS

21,773 views • 1 month ago

🇺🇸 ZUCK: META AI IS THE BIGGEST, BEST, IN THE WORLD!! Has anyone ever used it?... “Meta AI now has nearly 600 million monthly actives and as promised is on track to be the most used AI assistant in the world by the end of the year. Llama 3.3 is a new 70 billion parameter text model that performs about as well as our 405 billion parameter model but now it is easier and more efficient to run. So that is the last Lama 3.0 release. The next stop is Lama 4.0.” Source: Instagram

🇺🇸 ZUCK: META AI IS THE BIGGEST, BEST, IN THE WORLD!! Has anyone ever used it?... “Meta AI now has nearly 600 million monthly actives and as promised is on track to be the most used AI assistant in the world by the end of the year. Llama 3.3 is a new 70 billion parameter text model that performs about as well as our 405 billion parameter model but now it is easier and more efficient to run. So that is the last Lama 3.0 release. The next stop is Lama 4.0.” Source: Instagram

Mario Nawfal

482,099 views • 1 year ago

“Intelligence is being scaled. Care is not.” Geoffrey Hinton says the AI race is optimizing for intelligence, not for the kind of beings we’re creating. If labs are chasing IPOs and trillion-dollar valuations, the obvious move is to scale compute and ship faster. But smarter systems are not automatically systems that care about us.

“Intelligence is being scaled. Care is not.” Geoffrey Hinton says the AI race is optimizing for intelligence, not for the kind of beings we’re creating. If labs are chasing IPOs and trillion-dollar valuations, the obvious move is to scale compute and ship faster. But smarter systems are not automatically systems that care about us.

Jon Hernandez

15,146 views • 2 months ago