Threshold Crossings as Tail Events for Catastrophic AI Risk
Fuente:
arXiv
Enregistré dans:
| Auteur principal: | |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866917968328261632 |
|---|---|
| author | Perrier, Elija |
| author_facet | Perrier, Elija |
| contents | We analyse circumstances in which bifurcation-driven jumps in AI systems are associated with emergent heavy-tailed outcome distributions. By analysing how a control parameter's random fluctuations near a catastrophic threshold generate extreme outcomes, we demonstrate in what circumstances the probability of a sudden, large-scale, transition aligns closely with the tail probability of the resulting damage distribution. Our results contribute to research in monitoring, mitigation and control of AI systems when seeking to manage potentially catastrophic AI risk. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_18979 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Threshold Crossings as Tail Events for Catastrophic AI Risk Perrier, Elija Computers and Society Artificial Intelligence We analyse circumstances in which bifurcation-driven jumps in AI systems are associated with emergent heavy-tailed outcome distributions. By analysing how a control parameter's random fluctuations near a catastrophic threshold generate extreme outcomes, we demonstrate in what circumstances the probability of a sudden, large-scale, transition aligns closely with the tail probability of the resulting damage distribution. Our results contribute to research in monitoring, mitigation and control of AI systems when seeking to manage potentially catastrophic AI risk. |
| title | Threshold Crossings as Tail Events for Catastrophic AI Risk |
| topic | Computers and Society Artificial Intelligence |
| url | https://arxiv.org/abs/2503.18979 |