Thanks, Stephen! BTW, the last version of the paper is here and open access: link.springer.com/article/10.1...

I recommend reading that

Huge congratulations, Aylin!!!

Great work led by @andyliu.bsky.social and collaborators:
@kghate.bsky.social, @monadiab77.bsky.social, @daniel-fried.bsky.social, @atoosakz.bsky.social
Preprint: www.arxiv.org/abs/2509.25369

Kasirzadeh also has a really insightful, heartbreaking thread about what "normal" hides here: bsky.app/profile/atoo...

Anyway, big fan of the 3rd option Sigal presents, @atoosakz.bsky.social 's warning of "gradual accumulation of smaller, seemingly non-existential, AI risks" until catastrophe.

A warning that suggests AI safetyists should take AI ethics & sociology much more seriously! arxiv.org/pdf/2401.07836

We are very excited to have you, Bálint!

You can read AI as normal technology here: knightcolumbia.org/content/ai-a...

But because judgments of normal are too subjective—and because the world we live in entangled with advanved AI systems is anything but normal—the metaphor can conceals more than it reveals. So here is my proposal: Keep the demystification project live; drop the label “normal”. 10/n

—I’ve pushed back against the “super-intelligence will kill us all” narrative myself for years.
(See for example: Two types of AI existential risk: decisive and accumulative: link.springer.com/article/10.1... AI safety for everyone: www.nature.com/articles/s42...) 9/n

A system that decides who lives or dies is nothing like a household appliance. I admire the scholars Arvind Narayanan and Sayash Kapoor who coined the phrase and share their wish to redefine the notion of “normal” to move the AI governance debate beyond extreme doom. 8/n

The metaphor of "AI as normal technology" appeals cause it demystifies in the first place. But the label normal collapses the moment AI directs drones, misidentifies civilians, rewrites battlefield doctrine, or nudges users toward self-harm. 7/n

Seat belts were fussy in 1960, masks were not normal in 2019; both now signal basic care. Those who control information circulation or policy can freeze and stabilize a narrative in place, branding a live hazard normal by sheer repetition. 6/n

Normal isn't a fact; it's a perception built from repetition & power. My brother can call a night of missile fire normal simply because he's used to it.To an outsider, the scene is clearly abnormal. Judgements of normality rest on thin, shifting ground. History shows how quickly baselines move. 5/n

A toaster browns bread & stops; an agentic chatbot can keep learning and might suddenly write ransomware. Consumer devices have fixed functions & well-tested safety rules. Self-modifying agentic systems don't. Using the same label 4 both narrows the questions we ask about oversight & liability. 4/n