Self-awareness isn’t even the biggest issue. I teach the computational neuroscience of consciousness as a consequence of self-modeling to facilitate homeostasis. The human brain evolved these systems to manage embodied existence. Modern AI doesn’t need this. The potential threats are weirder.
M. C. Flux PhD
@fluxinflux.bsky.social
555 Followers
21 Following
Neuroscience and Clinical Psychology PhD
Educator and researcher. My research explores, equity, novel therapeutics, and human emotion. I’m passionate about communication and using information to empower communities.
Proudly queer and neurodiverse
-
-
The stupidity fools us. We are not dealing with Skynet, yes, but what we are dealing with holds dangers that are challenging to even imagine.
This is something we discuss a lot in my AI reading group (why I was reading that paper.) The computational architecture of our AI systems makes this weird.
-
Seriously. This is all becoming far too real.
-
Currently reading an article about the impact of AI hype. There’s so much here, but the impact of AI investment on the balance of military investment and power is the most damning.
(Highlights my own)
Paper: dl.acm.org/doi/full/10....
-
Thanks for the tag though!
-
This is my field! But I am not looking for a postdoc position 😅
-
Hi friend!!
-
I guess I’ve just cultivated my appearance to look surreal? Feel free to check me out! I am very much a real person 😅
-
Hahaha it’s doing its job then!
-
Fair! This is a clip from a lecture for my general psychology class where we discuss different biases. So it’s new for many of them.
![A screenshot from an academic paper about the impact of AI hype that is highlighted. It reads:
We see this dynamic at play in OpenAI's shifting stance and policies as they have become more closely tied to Microsoft, a choice they made admittedly to access scarce computational resources. This is illustrated by the 2024 change to their acceptable use policy, which significantly softened the proscription against using their products for "military and warfare"; this change unlocked new revenue streams [112], connected OpenAI's large scale AI to Microsoft's existing US military partnerships [135], and later pushed AI to direct battlefield application via a partnership with the defense-tech company Anduril [85]. The public did not have a say in this determi-nation, nor, we assume, did the global stakeholders whose interests may not be safeguarded by the militaries that OpenAl chooses to provide services to. And yet, the concentrated private industry power over Al creates a small, and financially incentivized segment of AI decision makers. We should consider how such concentrated power with agency over centralized Al could shape society under more authoritarian conditions.](https://cdn.bsky.app/img/feed_fullsize/plain/did:plc:bep462xlz63fm6a3lnmtgteb/bafkreia3mqazleigcyaejrcp7wvmj3ihmn2w2zccp2wufux736okp4cfha@jpeg)