works on
multivariate analysis 1non-normal dynamics 1signal processing 1time series inference 1transient amplification 1

From the 1 of 17 linked papers with an AI index.

collaborators
Showing cs.AIShow all

1 paper · 1 filter

cs.AI2026

Why AI Alignment Failure Is Structural: Learned Human Interaction Structures and AGI as an Endogenous Evolutionary Shock

Didier Sornette, Sandro Claudio Lera, Ke Wu

Recent reports of large language models (LLMs) exhibiting behaviors such as deception, threats, or blackmail are often interpreted as evidence of alignment failure or emergent mali…