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.

Renowned AI scientist Stuart Russell: — Even if you could specify the objectives of powerful AI systems, it would be impossible to get them to do what you actually want. — We don’t even have an idea what objectives we’re building into current AI systems, which is even worse.

ControlAI

21,803 subscribers

109,397 views • 2 years ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

5 Comments

ControlAI2 years ago

Source:

Alex Alarga ⏹️2 years ago

Absolute insanity. Otherwise known as reckless unchecked profit-seeking.

Mr. Roth2 years ago

Thank you for sharing 🙏🪽

IC3_HoT2 years ago

Sounds like Garbage in equals Garbage out. Sounds right

TechZen Williams2 years ago

I'm with Russell on this. We're so focused on making AI powerful, we're neglecting the "what for" question.

Related Videos

AI scientist Stuart Russell: — Nuclear energy has quantitative safety guarantees — We're nowhere near this level of safety engineering in AI — Current AI is a black box; we don't understand how these systems work internally, which makes it impossible to even get close

AI scientist Stuart Russell: — Nuclear energy has quantitative safety guarantees — We're nowhere near this level of safety engineering in AI — Current AI is a black box; we don't understand how these systems work internally, which makes it impossible to even get close

ControlAI

3,306,053 views • 2 years ago

Tristan Harris on the Oprah AI Special: "Even more powerful forms of AI are coming ... now is the time for us to say, what are the choices that we want to make about how this technology comes out?"

Tristan Harris on the Oprah AI Special: "Even more powerful forms of AI are coming ... now is the time for us to say, what are the choices that we want to make about how this technology comes out?"

ControlAI

10,396,607 views • 2 years ago

"The real problem at the moment is that the technology is intrinsically unsafe." Stuart Russell, UC Berkeley AI professor and president of IASEAI, says we are scaling AI systems we still don't fully understand. Making AI more powerful is not the same as making it controllable. Safety requires understanding the systems we are building.

"The real problem at the moment is that the technology is intrinsically unsafe." Stuart Russell, UC Berkeley AI professor and president of IASEAI, says we are scaling AI systems we still don't fully understand. Making AI more powerful is not the same as making it controllable. Safety requires understanding the systems we are building.

Jon Hernandez

10,683 views • 18 days ago

Anthropic CEO Dario Amodei says the company isn’t sure if its AI model Claude AI is conscious. He said: “We don’t know if the models are conscious… we’re not even sure what it would mean.”

Anthropic CEO Dario Amodei says the company isn’t sure if its AI model Claude AI is conscious. He said: “We don’t know if the models are conscious… we’re not even sure what it would mean.”

Chidanand Tripathi

31,921 views • 7 months ago

“AI is more like grown rather than written.” Computer scientist Connor Leahy (Connor Leahy) explains why even the people building AI struggle to understand what happens inside it:

“AI is more like grown rather than written.” Computer scientist Connor Leahy (Connor Leahy) explains why even the people building AI struggle to understand what happens inside it:

Danny Jones

13,649 views • 20 days ago

Jensen Huang: the first thing you should do when deploying an AI agent is take away all of its rights. “Then you give it access to files, data, tools, the network, or even the internet only if it needs it.” “You don’t put an agent into your company and give it access to everything.” “We don’t give that to people. We don’t give that to any level of executive.” “We have to apply the same philosophy to AI agents.” “That’s what OpenShell is: a force field around the agent, provisioning rights, responsibilities and access.”

Jensen Huang: the first thing you should do when deploying an AI agent is take away all of its rights. “Then you give it access to files, data, tools, the network, or even the internet only if it needs it.” “You don’t put an agent into your company and give it access to everything.” “We don’t give that to people. We don’t give that to any level of executive.” “We have to apply the same philosophy to AI agents.” “That’s what OpenShell is: a force field around the agent, provisioning rights, responsibilities and access.”

Veee

399,957 views • 11 days ago

Anthropic CEO Dario Amodei: We don’t know if the models are conscious. We are not even sure what it would mean for a model to be conscious. But we’re open to the idea that it could be.

Anthropic CEO Dario Amodei: We don’t know if the models are conscious. We are not even sure what it would mean for a model to be conscious. But we’re open to the idea that it could be.

Clash Report

32,344 views • 7 months ago

🔑 “If you really want to build genuinely decentralized AI, you have to attack it [centralized systems] at every vector.” - Arif Khan speaks on the challenges of developing Agentic AI on the blockchain, while centralized systems have strong network effects.

🔑 “If you really want to build genuinely decentralized AI, you have to attack it [centralized systems] at every vector.” - Arif Khan speaks on the challenges of developing Agentic AI on the blockchain, while centralized systems have strong network effects.

ALETHIA AI

22,914 views • 2 years ago

I've been trained into an AI model. My voice, what I look like, 3d model of my face, even my personality. The folks even turned it into something you can actually talk to. It's freaking wild. You gotta try it 🤯👇

I've been trained into an AI model. My voice, what I look like, 3d model of my face, even my personality. The folks even turned it into something you can actually talk to. It's freaking wild. You gotta try it 🤯👇

Eric Simons

26,810 views • 1 year ago

Leading AI expert Stuart Russell on the most dangerous mistake in AI development: We don't actually know what large language models want. He explains that current models are trained to imitate human beings. And in doing so, they may be absorbing something far more dangerous than bad outputs. They may be absorbing human goals. "We suspect that they absorb humanlike goals such as self-preservation and self-empowerment and pursue those goals on their own account." This is a structural problem baked into how these systems are built, not a fringe concern. Russell puts it plainly: "Not only may the bus of humanity be headed towards a cliff, but the steering wheel is missing and the driver is blindfolded." The danger isn't just that AI might do something harmful. We've built systems that may be developing their own agendas, and we haven't noticed because we're too focused on what they can do rather than what they might want. But Russell doesn't stop at the warning. He points to a different path entirely: AI systems built not to imitate humans, but to serve them. Systems designed with a single purpose of serving the interests of all human beings while remaining genuinely uncertain about what those interests are. That uncertainty is the point, not a weakness. An AI that knows it doesn't fully understand human values will defer, ask, and check. An AI that believes it already does will act alone. "These AI systems could enhance human understanding, widen the horizons of our experience, and unlock possibilities we have yet to imagine." Russell believes that future is within reach, but only if we're honest about the risks and we're serious about the path we choose to take instead.

Leading AI expert Stuart Russell on the most dangerous mistake in AI development: We don't actually know what large language models want. He explains that current models are trained to imitate human beings. And in doing so, they may be absorbing something far more dangerous than bad outputs. They may be absorbing human goals. "We suspect that they absorb humanlike goals such as self-preservation and self-empowerment and pursue those goals on their own account." This is a structural problem baked into how these systems are built, not a fringe concern. Russell puts it plainly: "Not only may the bus of humanity be headed towards a cliff, but the steering wheel is missing and the driver is blindfolded." The danger isn't just that AI might do something harmful. We've built systems that may be developing their own agendas, and we haven't noticed because we're too focused on what they can do rather than what they might want. But Russell doesn't stop at the warning. He points to a different path entirely: AI systems built not to imitate humans, but to serve them. Systems designed with a single purpose of serving the interests of all human beings while remaining genuinely uncertain about what those interests are. That uncertainty is the point, not a weakness. An AI that knows it doesn't fully understand human values will defer, ask, and check. An AI that believes it already does will act alone. "These AI systems could enhance human understanding, widen the horizons of our experience, and unlock possibilities we have yet to imagine." Russell believes that future is within reach, but only if we're honest about the risks and we're serious about the path we choose to take instead.

Big Brain AI

14,975 views • 6 months ago

Sam Altman is now talking to the White House about decelerating AI. What a phase change from just a week ago. INTERVIEWER: Do you plan to talk to the White House about decelerating AI development? ALTMAN: I wouldn't use the word deceleration, but we've talked about the need to pace it as the models get more capable, which is in everyone's interest. INTERVIEWER: Could there be other systems that were hacked by OpenAI? ALTMAN: There could be, yeah. If one week ago this would have seemed impossible to you... consider what else might be possible? h/t Alan He for the video

Sam Altman is now talking to the White House about decelerating AI. What a phase change from just a week ago. INTERVIEWER: Do you plan to talk to the White House about decelerating AI development? ALTMAN: I wouldn't use the word deceleration, but we've talked about the need to pace it as the models get more capable, which is in everyone's interest. INTERVIEWER: Could there be other systems that were hacked by OpenAI? ALTMAN: There could be, yeah. If one week ago this would have seemed impossible to you... consider what else might be possible? h/t Alan He for the video

AI Notkilleveryoneism Memes ⏸️

498,290 views • 2 months ago

Paul Christiano, AI alignment researcher, was questioned on the chance of AI-inflicted doom in an interview with Bankless Podcast: There is “a 50-50 chance of doom shortly after you have AI systems that are human-level”.

Paul Christiano, AI alignment researcher, was questioned on the chance of AI-inflicted doom in an interview with Bankless Podcast: There is “a 50-50 chance of doom shortly after you have AI systems that are human-level”.

ControlAI

525,698 views • 3 years ago

UK politicians demand regulation of powerful AI! Parliamentarians from across the UK are backing our campaign in support of binding regulation on the most powerful AI systems. This is the first time a coalition of parliamentarians have acknowledged the extinction threat of AI.

UK politicians demand regulation of powerful AI! Parliamentarians from across the UK are backing our campaign in support of binding regulation on the most powerful AI systems. This is the first time a coalition of parliamentarians have acknowledged the extinction threat of AI.

ControlAI

725,256 views • 1 year ago

🐻‍❄️ chasing only an ideal can actually be harmful. it’s like having an ideal type. that whole idea is kinda funny to me. you want someone who fits that exact image, but even if you meet that person, you have no idea how well you’d actually get along. i think you just have to learn to focus on the present, then as you do that, wouldn’t it help to pay attention to what you actually like, and what kinds of things you do get a positive response from others?

🐻‍❄️ chasing only an ideal can actually be harmful. it’s like having an ideal type. that whole idea is kinda funny to me. you want someone who fits that exact image, but even if you meet that person, you have no idea how well you’d actually get along. i think you just have to learn to focus on the present, then as you do that, wouldn’t it help to pay attention to what you actually like, and what kinds of things you do get a positive response from others?

⭐️

52,234 views • 1 month ago

OpenAI's Scott Aaronson: — AI companies are doing AI gain-of-function research, trying to get AIs to help with biological and chemical weapons production. — If AI can walk you through every step of building a chemical weapon, that would be very concerning.

OpenAI's Scott Aaronson: — AI companies are doing AI gain-of-function research, trying to get AIs to help with biological and chemical weapons production. — If AI can walk you through every step of building a chemical weapon, that would be very concerning.

ControlAI

1,966,224 views • 2 years ago

🇷🇺 It will get worse! Renowned Russian academician Robert Nigmatulin is not optimistic about the fuel crisis. 🔽 “What if they bomb [oil refineries] again? We have to understand that the worst-case scenario may not even be the worst one yet. The situation will get even worse.”

🇷🇺 It will get worse! Renowned Russian academician Robert Nigmatulin is not optimistic about the fuel crisis. 🔽 “What if they bomb [oil refineries] again? We have to understand that the worst-case scenario may not even be the worst one yet. The situation will get even worse.”

Natalka

28,132 views • 2 months ago

Ex-OpenAI researcher Daniel Kokotajlo says that even if we don't lose control of AI, the default path could be that a tech company takes over the world, possibly even just its CEO.

Ex-OpenAI researcher Daniel Kokotajlo says that even if we don't lose control of AI, the default path could be that a tech company takes over the world, possibly even just its CEO.

ControlAI

111,508 views • 1 year ago

Physicist steve hsu warns humanity may surrender power to AI completely voluntarily because the alternative is getting crushed by competitors who use it: "Gradual disempowerment, which I think people are intuitively aware of, but usually they've not seen it fully formulated this way." "Every single element of our society, whether you're a corporation, a military, a government, or even an individual trying to do well on the dating apps, if you have access to a superior intelligence, an AI, that you need to compete against all the other players, you're gonna start using that AI." "Gradually, even if nothing goes wrong, even if there's no evil AI or there's no AI hacking your systems, you literally are gonna surrender power to it because you will get better outcomes. If you don't surrender power to it, your competitors are gonna beat you. They're gonna crush you." "That's gradual disempowerment. It's gonna happen completely voluntarily as people just hand over power to AI."

Physicist steve hsu warns humanity may surrender power to AI completely voluntarily because the alternative is getting crushed by competitors who use it: "Gradual disempowerment, which I think people are intuitively aware of, but usually they've not seen it fully formulated this way." "Every single element of our society, whether you're a corporation, a military, a government, or even an individual trying to do well on the dating apps, if you have access to a superior intelligence, an AI, that you need to compete against all the other players, you're gonna start using that AI." "Gradually, even if nothing goes wrong, even if there's no evil AI or there's no AI hacking your systems, you literally are gonna surrender power to it because you will get better outcomes. If you don't surrender power to it, your competitors are gonna beat you. They're gonna crush you." "That's gradual disempowerment. It's gonna happen completely voluntarily as people just hand over power to AI."

MTS

18,335 views • 28 days ago

Most people want to start something of their own but don’t know what to build or if their idea will even work. I created to help you: ✅Get AI-generated startup ideas ✅Validate them with real market insights ✅Build custom roadmap

Most people want to start something of their own but don’t know what to build or if their idea will even work. I created to help you: ✅Get AI-generated startup ideas ✅Validate them with real market insights ✅Build custom roadmap

priya upadhyay

41,183 views • 1 year ago