Postdoctoral Fellow in Philosophy of AI at Philosophy Department at Lingnan University and the HKCRC. Formerly: University of Cambridge, Department of History and Philosophy of Science. 🏳️🌈
https://scholar.google.com/citations?user=qlNi3AoAAAAJ&hl=en
David Jacob Harrison
Lastly, I also want to thank @twachter.bsky.social and @michaelgarfield.bsky.social for their invaluable input and feedback, and the @hkcrc.bsky.social and @divintelligence.bsky.social for being the hosting spaces where earlier versions of this idea were presented.
Though this is likely whatever stage comes two stages after late-stage capitalism.
I explore how the value landscape to which AI is being aligned is characterized by value capture and flattening. I suggest that this 'flattening' establishes a context in which AI alignment occurs - achieving alignment in the 'reverse direction' by humans adapting to machines.
The RAP is a terminological intervention and what I want to contribute to in developing this idea going forward is that alignment is instead of (or in addition to) an engineering problem of control it is also a socio-technical problem of power.
It feels like, at long last, the half-jocular, half-serious end point of the rhetorical question 'what stage of capitalism is this?' has finally been reached with this product.
gudtrip.com
Philosophically, I want to thank @kulveit.bsky.social and his colleagues for their work on 'gradual disempowerments' in the alignment debate. I especially am indebted to the work of @add-hawk.bsky.social whose notion of value capture is explored as one mechanism behind the RAP.
I'm very excited to announce my first Postdoctoral publication in AI & Society entitled 'The Message Hidden Within the Pattern: A Reverse Alignment Problem for Debates in Artificial Intelligence'.
link.springer.com/article/10.1...
I am thrilled by the engagement with my new paper in AI & Society on the Reverse Alignment Problem so far.
link.springer.com/article/10.1...
Conceptually, this work is indebted to the work of @cwarzel.bsky.social, @chaykak.bsky.social and @karenhao.bsky.social, among many others, whose reporting deeply informed my approach to the topic of alignment.
This paper explores what I call the reverse alignment problem (RAP). The alignment problem in artificial intelligence (AI) is the challenge of ensuring that superintelligent AI systems harmonize and p...
