LESSWRONG
LW

Wikitags

Threat Models (AI)

Edited by Quinn, Jacob Pfau, et al. last updated 12th Apr 2023

A threat model is a story of how a particular risk (e.g. AI) plays out.

In the AI risk case, according to Rohin Shah, a threat model is ideally:

Combination of a development model that says how we get AGI and a risk model that says how AGI leads to existential catastrophe.

See also AI Risk Concrete Stories

Subscribe
1
Subscribe
1
Discussion1
Discussion1
Posts tagged Threat Models (AI)
249Another (outer) alignment failure story
Ω
paulfchristiano
4y
Ω
38
437What failure looks like
Ω
paulfchristiano
6y
Ω
55
74Distinguishing AI takeover scenarios
Ω
Sam Clarke, Sammy Martin
4y
Ω
11
285What Multipolar Failure Looks Like, and Robust Agent-Agnostic Processes (RAAPs)
Ω
Andrew_Critch
4y
Ω
65
48Vignettes Workshop (AI Impacts)
Ω
Daniel Kokotajlo
4y
Ω
6
949AGI Ruin: A List of Lethalities
Ω
Eliezer Yudkowsky
3y
Ω
712
315On how various plans miss the hard bits of the alignment challenge
Ω
So8res
3y
Ω
89
113Less Realistic Tales of Doom
Ω
Mark Xu
4y
Ω
13
65Survey on AI existential risk scenarios
Ω
Sam Clarke, apc, Jonas Schuett
4y
Ω
11
31Investigating AI Takeover Scenarios
Ω
Sammy Martin
4y
Ω
1
370Without specific countermeasures, the easiest path to transformative AI likely leads to AI takeover
Ω
Ajeya Cotra
3y
Ω
95
132Current AIs Provide Nearly No Data Relevant to AGI Alignment
Ω
Thane Ruthenis
2y
Ω
157
71Rogue AGI Embodies Valuable Intellectual Property
Ω
Mark Xu, CarlShulman
4y
Ω
9
170Will the growing deer prion epidemic spread to humans? Why not?
eukaryote
2y
33
170AI Could Defeat All Of Us Combined
HoldenKarnofsky
3y
42
Load More (15/86)
Add Posts