What is your biggest concern about AI safety / alignment?
I have some fear of super intelligent AGI foom stuff, but I also think the near-term small AI stuff will cause a huge upheaval in the distribution of power and not necessarily for the good
Conversation
couple people have brought this up, I think we need a term for this. I propose "machinewashing"
twitter.com/evilJBP/status
This Tweet is unavailable.
3
1
42
another common concern: the machines make predictions based on their data sets, the world changes based on their predictions, the biases of the data are reinforced
Quote Tweet
Replying to @deepfates
in general, wild nth order effects of unexpected feedback loops
4
35
many people are concerned with the way those in power will use even more power. Ted brings another point: what about the people without power?
Butlerian Jihad anybody?
Quote Tweet
Replying to @deepfates
β’ zero fear of foom
β’ very small (0.2%) fear about "alignment" of systems
β’ 55% fear about the alignment of large corporations & state security apparatuses using ML
β’ 44.8% fear about the politics of underemployed workers led by crusading "tech critics" on both right and left
3
26
Yes - the reaction will make the populist movements of the early 20th c. look like weekend larps



