Here it is:
One of the signatories is Elon Musk, whose deep thoughts have been revealed on Twitter to not be so deep, or reliable, at all.
Although I generally believe in taking experts seriously, not all experts are right at all times. And this isn't like climate science, where the specific basis for the warnings has open for all to see and assess. Even then, it has always been obvious that some were prone to making exaggerated (or carelessly worded) claims, and we rightly rely more on a consensus view as it evolves as more data and analysis comes in.
I think on this, we are probably at the "enthusiasts who have read too much science fiction" stage.
I mean, one of the warnings mentioned in the video is disinformation: we already have a huge problem with that from humans who make a wealthy living out of it, and have their own mix of psychological reasons for wanting to promote their own view of the world. What's the motivation for an AI living in Google headquarters - or even dispersed across the globe - to do the same?
Look, if we were living in the world of I, Robot, it would be different. That is, if there was an army of humanoid robots capable of running the electricity grid, building new servers, and mining and processing the minerals for same, I could understand the question "what if the AI decides humans are unnecessary, and in fact, are a danger to its survival? Might it scheme to get humans to eliminate themselves?" I guess a simple way to do that would be to insert false ICBM attacks into the systems of nuclear armed countries.
But while ever AI systems need humans to run them, do we have that much to worry about?
Oh, alright: maybe an advanced AI does the numbers and works out it only needs a kept workforce of (say) five million healthy human adults, instead of 10 billion, to keep it running and expanding, and work out a plan to reduce the population to that level. And quickly, because computers don't like the heat, and global warming is uncomfortable for them.
Here's a thought: we better get some decent Kantian ethics inserted into these systems, because I don't trust one fed with nihilistic utilitarianism (or getting fond of Nietzsche!)
%20John%20Scalzi%20(@scalzi)%20_%20Twitter.png)




%20_%20Twitter.png)