I’d push back on this a bit. It’s not the agents that will miss the morality and context … it’s us, if we build them that way.
AI systems don’t have values of their own. They reflect the intent, blind spots, and imagination of the people who create and deploy them. If we treat AI as something independent, detached from human judgment, we’re already giving up our role in shaping it. The future doesn’t have to be contextless. But it will be if we keep designing tech optimized for scale, speed, and efficiency … without asking what kind of world we want to live in.
I guess ultimately AI is still a reflection of the data we feed it, which essentially is a mirror to what the representation of humanity in the data sets we present to it—with all its imperfections, biases, and all.
Then perhaps the real consideration is should we worry that the data we are feeding it is an imperfect representation of humanity, in regards to the fact that we are still in progress to correcting the flaws we are only starting to uncover from the previous Internet boom.
Or do you have any suggestions on how we can feed AI or build safer and more fair AI systems for the future?
I’d push back on this a bit. It’s not the agents that will miss the morality and context … it’s us, if we build them that way.
AI systems don’t have values of their own. They reflect the intent, blind spots, and imagination of the people who create and deploy them. If we treat AI as something independent, detached from human judgment, we’re already giving up our role in shaping it. The future doesn’t have to be contextless. But it will be if we keep designing tech optimized for scale, speed, and efficiency … without asking what kind of world we want to live in.
Thats a fair point.
I guess ultimately AI is still a reflection of the data we feed it, which essentially is a mirror to what the representation of humanity in the data sets we present to it—with all its imperfections, biases, and all.
Then perhaps the real consideration is should we worry that the data we are feeding it is an imperfect representation of humanity, in regards to the fact that we are still in progress to correcting the flaws we are only starting to uncover from the previous Internet boom.
Or do you have any suggestions on how we can feed AI or build safer and more fair AI systems for the future?
To add as well, it’s the unintended consequences that we’re not seeing, a decade and beyond that I’m most concerned about.
Just like how we didn’t know how social media and algorithms can affect the world as it happens in real time today.