Compassionate A.I.
Let me be very positive, envisioning a future in which A.I. will have as the main characteristic that which also is the best in human being.
Let me call it ‘Compassion.’
A term does not have the magical feature of conjuring up a concept unless this concept is well defined and agreed upon. To me and in this text, Compassion – note the big C – is very close to ‘empathy’ [see: “Landscape of Empathy”].
Main difference: whereas empathy is utterly concrete, Compassion is more like an ideal that one can strive for, be attracted towards.
As for instance in Buddhism
In a historical as well as essential sense, Compassion is more important in (at least my view upon) Buddhism than the Buddha himself. Somehow, of course, they perfectly blend.
In the other great religions and worldviews too, Compassion plays a major role, showing that in the human scheme of values, it has always been of prime significance
and is and will always be.
Basic aim of morality
Compassion has two sides: to relieve suffering and to attain human growth [see: “Two-Sided Compassion”].
Bringing these two together, seeing them as one, is a very ethical stance. To me, they both together form the basic aim of morality.
My hope is that the future will be one in which this can be realized to a huge degree. Since A.I. will be a very important part of the future, we also need to think about this, without delay.
Doing so is a Compassionate endeavor.
Next door, A.I.
We shouldn’t be coy about this. A.I. is behind next door, whether talking in years or a bit longer.
According to me, there is only one possible humane way in which it can come through that door. This way can be called Compassion.
I am personally very hopeful that this will indeed be the way, eventually. For the first stages of it, I’m more apprehensive. At the same time, these are the stages in which we, humans, bear most responsibility. Right now, even.
The time to value-align is now.
Aligning our values between ourselves, as human beings.
Aligning our values as human beings, to those of A.I.
Compassion.
In this text, as a human, I’ve put humanity central as should be: human growth, a humane way… Without losing any of this, Compassion is transcending. It looks at all sentient beings, including animals, including A.I. itself.
According to me, it is within this that also as a species lies our brightest future.
Related Posts
How Lisa Prevents LLM Hallucinations
Hallucinations (better-called confabulations) in the context of large language models (LLMs) occur when these models generate information that isn’t factually accurate. Lisa can mitigate these from the insight of why they happen, namely: LLM confabulations happen because these systems don’t have a proper understanding of the world but generate text based on patterns learned from Read the full article…
From Analogy to Intelligence
If an inference system (mind or A.I.) holds many patterns, it may find many similarities between these at several levels of abstraction. The human brain is an immense pattern recognizer. Super-A.I. will undoubtedly follow a similar path ― as does Lisa. Lisa’s journey thus exemplifies the potential of A.I. to evolve from mere data processors to entities capable Read the full article…
The Compute-Efficient Frontier
Research on scaling laws for LLMs suggests that while scaling model size (functional parameters), dataset size (tokens), and compute (amount of computing power) improves performance, diminishing returns are becoming evident. This is called the ‘compute-efficient frontier.’ Apparently, it does not depend much on architectural details (such as network width or depth) as long as reasonably Read the full article…