I want to say this again…AI may be great, but it can’t be greater than a living being. I have worked and experimented a lot with AI (1,2,3,4) over the past years, and have been updating my perspectives to better understand its potential, flaws, where it shines or otherwise.
Is AI a wonderful application that has helped make a lot of processes in coding, tech, search, research better? Yes. Is it great and a truly great appliance the same way as invention of electricity and internet and a blender? Yes. Is it a tool the same way a compiler is? Yes.
I feel like there are so many confusing the abilities with that of a sentient being, rather than a machine that powers down if you unplug, because it mimics the nature of that of a human thought - the same way a pond mimics the reflection but doesn’t have the same organs 5.
This topic came up recently as I was discussing few of the ideas with a person and they were saying how a lot of philanthropic grants are towards AI safety / Global ‘catastrophic risk’ based on a statistical calculation that “even a tiny, sub-1% reduction in existential risk yields an expected value in the billions or trillions of lives saved” and we are talking about population centuries into the future.
I do believe that this “Expected Value = P(extinction) x Number of future lives lost” is more of doomsday narrative pov that makes the folks and companies building AI to feel self important. (You are so worthy of all the love and goodness regardless of anything you create, folks. I respect the living beings as I started this with, and this is about the perspective and idea, not any humans.)
I was told that as ‘someone who uses AI a lot in work,’ if I worked on AI alignment and safety-focused themes, they would be more than willing to pay 10x the fee I proposed because apparently focusing on a ‘few hundreds’ of living beings is ‘less valuable’ than pouring everything into the chance that AI becomes a superbeing that suddenly outmaneuvers all of humanity, rather than a tool that shuts off when unplugged.
I find it really disturbing, disappointing and sad. So, no please don’t email me proposing any AI alignment to ‘prevent global catastrophic risk’ (it violates #2 and #3 in the criteria if you want me to be specific), I would rather work on problems that are much more unglamarous but real to the current beings. Thank you.
P.s - I love sci-fi novels, btw. Just that I wish the money in real could move to underfunded and necessary problems that humans are currently facing instead of a narrative that helps with marketing.