So, picture it: A civilisation past its peak, that’s invented AI but can no longer build it, relying on ancient black boxes to tell it what to do. Nobody understands how the box knows, but experience suggests that it’s better to follow its advice.
Originally shared by Singularity Hub
Life-or-Death Algorithms: The Black Box of AI in Medicine, and How to Avoid It
Like Logan’s Run?
Like Logan’s Run?
I’m only vaguely familiar with that, but probably.
I’m only vaguely familiar with that, but probably.
Here’s an idea I’ve had for a while: instead of programming AI or even ML to “protect humans” and such, make the directive “teach humans to fulfill their potential both individually and severally.”
That way we become the best we can be, led by both the machines and fellow humans – because some humans will want to teach and the programs will help them become the best teachers they can be.
It becomes a cycle of innovation instead of replacing work and humanity stagnates.
Here’s an idea I’ve had for a while: instead of programming AI or even ML to “protect humans” and such, make the directive “teach humans to fulfill their potential both individually and severally.”
That way we become the best we can be, led by both the machines and fellow humans – because some humans will want to teach and the programs will help them become the best teachers they can be.
It becomes a cycle of innovation instead of replacing work and humanity stagnates.
I like that a lot.
I like that a lot.
Mike Reeves-McMillan it’s a spin-off of an idea I had for a product I never learned enough to make. It was to monitor not only student progress, but also track how each student learned and find which learning paths resulted in the best results (defined by retention over time and application to higher level work.)
Mike Reeves-McMillan it’s a spin-off of an idea I had for a product I never learned enough to make. It was to monitor not only student progress, but also track how each student learned and find which learning paths resulted in the best results (defined by retention over time and application to higher level work.)
Samuel Smith Who determines what the potential of humans, both individually and severally, is?
Samuel Smith Who determines what the potential of humans, both individually and severally, is?
Valkyrie Page the humans. Good question!
Edit: humans determine direction, but we are often ignorant of how far we can go.
Valkyrie Page the humans. Good question!
Edit: humans determine direction, but we are often ignorant of how far we can go.
Samuel Smith Of course, the other important question is who determines the purpose? People, individually and severally, might have a variety of potentials. Some might have a potential as an assassin as well as an artist. How would AI help develop a potential artist? Presumably, not all humans would succeed in the direction in which they were pointed. We are human, after all. And what would be the overall goal? What would the desired society look like, and who would be making that choice…?
Further, might this not take some of the serendipity and inspired accidents/inspirations out of life?
Samuel Smith Of course, the other important question is who determines the purpose? People, individually and severally, might have a variety of potentials. Some might have a potential as an assassin as well as an artist. How would AI help develop a potential artist? Presumably, not all humans would succeed in the direction in which they were pointed. We are human, after all. And what would be the overall goal? What would the desired society look like, and who would be making that choice…?
Further, might this not take some of the serendipity and inspired accidents/inspirations out of life?