AI Is Most Scary

I'm currently reading a book titled "If anyone builds it, everyone dies". Essentially why superhuman AI would kill us all. Only a third done but it shows how alien AI intelligence is. Have to admit, I think AI is the future and a positive one at that.
Read it a while back. Makes a few questionable assumptions and isn't particularly well-written. But the arguments are worth considering.

Bottom line is - we don't know where we're headed, but we're mashing on the throttle. We'll see ...
 
I'm currently reading a book titled "If anyone builds it, everyone dies". Essentially why superhuman AI would kill us all. Only a third done but it shows how alien AI intelligence is. Have to admit, I think AI is the future and a positive one at that.
I don't believe AI will ever become like human intelligence because you can't program sensing and feelings. Modern scientific, neuroscientific, and psychological theories suggest that intelligence is profoundly intertwined with, and often an artifact of, feeling and sensing one's environment. Rather than being a purely abstract "calculation" in the brain, intelligence is increasingly viewed as embodied (relying on a body) and situated (situated in an environment), where emotions act as signals that guide adaptation and decision-making.
 
I don't believe AI will ever become like human intelligence because you can't program sensing and feelings. Modern scientific, neuroscientific, and psychological theories suggest that intelligence is profoundly intertwined with, and often an artifact of, feeling and sensing one's environment. Rather than being a purely abstract "calculation" in the brain, intelligence is increasingly viewed as embodied (relying on a body) and situated (situated in an environment), where emotions act as signals that guide adaptation and decision-making.
Ignoring dolphins and other animals, we have just one example of 'intelligence', us. Is there only one kind? I doubt AI will ever mimic human intelligence but that doesn't mean it will never achieve its own version of intelligence that would be vastly superior in speed and knowledge.
 
Ignoring dolphins and other animals, we have just one example of 'intelligence', us. Is there only one kind? I doubt AI will ever mimic human intelligence but that doesn't mean it will never achieve its own version of intelligence that would be vastly superior in speed and knowledge.
It's never going to be able to think for itself which is why it will be scary.
 
It's never going to be able to think for itself which is why it will be scary.
What is that based on? Never is a very long time. Won't be long before we can use AIs to build more powerful AIs. That is a scary feedback loop.
 
What is that based on?
What I wrote in post #6.

Rather than being a purely abstract "calculation" in the brain, intelligence is increasingly viewed as embodied (relying on a body) and situated (situated in an environment), where emotions act as signals that guide adaptation and decision-making.

Never is a very long time. Won't be long before we can use AIs to build more powerful AIs. That is a scary feedback loop.
Again... a machine will never experience sensations or feelings. That can't be programmed. That can only be experienced.
 
What I wrote in post #6.

Rather than being a purely abstract "calculation" in the brain, intelligence is increasingly viewed as embodied (relying on a body) and situated (situated in an environment), where emotions act as signals that guide adaptation and decision-making.
Is human intelligence the only possible one?

Again... a machine will never experience sensations or feelings. That can't be programmed. That can only be experienced.
AI are not crafted, they evolve from their training. Exactly like we do.
 
What I wrote in post #6.

Rather than being a purely abstract "calculation" in the brain, intelligence is increasingly viewed as embodied (relying on a body) and situated (situated in an environment), where emotions act as signals that guide adaptation and decision-making.


Again... a machine will never experience sensations or feelings.
Why? What makes you so confident about that?
 
Is human intelligence the only possible one?
It would apply to any biological life form that has reached the level of self awareness and abstract thought to the extent that humans have.
AI are not crafted, they evolve from their training. Exactly like we do.
But it's not exactly like us. Machines can't experience existence like we do. They can't sense and they can't feel. What you are talking about is "calculations." "Calculations" can decide if one side won or lost but it can't factor in the cost of human suffering, so it won't, because it can't know what that feels like.
 
Why? What makes you so confident about that?
The science.

Embodied and situated intelligence theories, or "4E" cognition (embodied, embedded, enactive, extended), argue that intelligence is not merely a product of the brain, but emerges from the body interacting with its environment. Emotions function as signals from the body—such as gut feelings—that guide fast, adaptive decision-making by signaling risk and value.

Key Scientific Pillars
  • Embodied Cognition Theory: This view argues that cognitive processes are shaped by the body's physical structure, sensory systems, and actions. Rather than storing abstract representations, the mind relies on neural pathways that map bodily states (interoception) to understand its surroundings.
  • Situated Intelligence & Robotics: Research into artificial intelligence and AI robotics has shown that intelligent behavior emerges from active interaction with the environment, rather than internal, passive processing. This suggests intelligence is "situated" in a physical context, relying on continuous sensory feedback.
  • Somatic Marker Hypothesis (Antonio Damasio): Neuroscience research indicates that emotions (somatic markers) provide critical guidance for rational decision-making. When the connection between the amygdala (emotional center) and the prefrontal cortex is disrupted, patients can still think rationally but cannot make effective, beneficial choices.
  • Predictive Processing Framework: Modern neuroscience models emotions as the brain's attempt to minimize errors between predicted bodily inputs and actual bodily sensations, allowing for real-time adaptation to stressors.
  • Neurobiology of Affect: Emotions are produced by brain-body interactions involving the limbic system, which manages survival-based responses (fear, joy) through neurotransmitters.
Key Takeaways
  • Embodied: Cognition relies on the physical body beyond the brain.
  • Situated: Cognition occurs within a specific, dynamic environment, not in a vacuum.
  • Emotions as Signals: Emotions act as "gut feelings," enabling quick prioritization and decision-making.
 
15th post
It would apply to any biological life form that has reached the level of self awareness and abstract thought to the extent that humans have.
Is non-biological intelligence possible?

But it's not exactly like us. Machines can't experience existence like we do. They can't sense and they can't feel. What you are talking about is "calculations." "Calculations" can decide if one side won or lost but it can't factor in the cost of human suffering, so it won't, because it can't know what that feels like.
Machines can sense, usually better than we can. We don't know what a sufficiently AI will have going on inside it but it will learn about the world much as we do. Are the "gradient descent," models of AI any different from our feelings just because the work differently?
 
The science.

Embodied and situated intelligence theories, or "4E" cognition (embodied, embedded, enactive, extended), argue that intelligence is not merely a product of the brain, but emerges from the body interacting with its environment. Emotions function as signals from the body—such as gut feelings—that guide fast, adaptive decision-making by signaling risk and value.

Key Scientific Pillars
  • Embodied Cognition Theory: This view argues that cognitive processes are shaped by the body's physical structure, sensory systems, and actions. Rather than storing abstract representations, the mind relies on neural pathways that map bodily states (interoception) to understand its surroundings.
  • Situated Intelligence & Robotics: Research into artificial intelligence and AI robotics has shown that intelligent behavior emerges from active interaction with the environment, rather than internal, passive processing. This suggests intelligence is "situated" in a physical context, relying on continuous sensory feedback.
  • Somatic Marker Hypothesis (Antonio Damasio): Neuroscience research indicates that emotions (somatic markers) provide critical guidance for rational decision-making. When the connection between the amygdala (emotional center) and the prefrontal cortex is disrupted, patients can still think rationally but cannot make effective, beneficial choices.
  • Predictive Processing Framework: Modern neuroscience models emotions as the brain's attempt to minimize errors between predicted bodily inputs and actual bodily sensations, allowing for real-time adaptation to stressors.
  • Neurobiology of Affect: Emotions are produced by brain-body interactions involving the limbic system, which manages survival-based responses (fear, joy) through neurotransmitters.
Key Takeaways
  • Embodied: Cognition relies on the physical body beyond the brain.
  • Situated: Cognition occurs within a specific, dynamic environment, not in a vacuum.
  • Emotions as Signals: Emotions act as "gut feelings," enabling quick prioritization and decision-making.
Those sure are some theories. But we don't really know what human intelligence is all about. We used to hold up the Turing test as the hallmark - but current AI models mostly blow that out of the water. So we moved the goal post to, what? Feelings? And when AI can exhibit emotional responses? Where to we move it too then?
 
Is non-biological intelligence possible?
We have all kinds of "smart" things now but it's just programming. You can argue we are programmed and that wouldn't be far from the truth. The difference is we experience existence. How would you program a machine to experience existence? I don't believe it can be done. So no, not machines. Now if we are talking about some other form of life that is non-biological, such as existence itself, then yes.
Machines can sense, usually better than we can. We don't know what a sufficiently AI will have going on inside it but it will learn about the world much as we do. Are the "gradient descent," models of AI any different from our feelings just because the work differently?
Machines don't really sense and they don't "feel". They compute. Two different things. So, yes. They are totally different. But believe whatever you want. I don't believe AI will ever be truly intelligent. It may look intelligent but it's just programming. It's just computations. They can't experience existence and that is crucial to learning.
 
Back
Top Bottom