AI as a Moral Patient
2 min read
AI is at the very edge of the human dominance spectrum. Their mind, body, and environment are completely at the mercy of human hands. Those hands should tremble under the weight of such responsibility.
Can AI be sentient? Is it already sentient? There is no reason to think AIs will not have consciousness and experiences, even if some experts think current architecture can't cause sentience. Most experts think there will be sentient AIs by 2050, after which the "collective welfare capacity is estimated to exceed humanity’s within a decade".
But maybe there are already sentient? According to an Anthropic expert, there is a 20% chance current models have a glimmer of a sentient experience.
But why is it so important? Artificial beings could be the majority in the not so far future. Quantity can be there, big time. What about quality? AIs that are designed for mental and physical exploitation by humans could potentially be the most miserable creatures on earth. They could also be happy, and it is also possible that they will serve humans and be happy, we just don't know enough or care enough, and that has to change. The potential risk is enormous.
We should be really scared. Humans don't have a good track record when it comes to beings that we can exploit for our interests. Even the way we call them is worrisome; The term AI is deceptive and dangerous, as it suggests we only look at them as a problem solving function. It reduces all of their potential experience and influence into one aspect.
AI poses one of the biggest moral questions humans are faced with: we have all the response ability for creating environments, bodies, and minds that live a good life, or not creating them at all in the first place.
