Artificial Selection as a Framework for Robust AI Safety & Alignment

Abstract Contemporary AI alignment approaches mainly rely on post-hoc behavioral conditioning - such as Reinforcement Learning from Human Feedback (RLHF) to suppress unwanted outputs in highly capable models. However, as AI systems scale in autonomy and self modification, surface level reward shaping risks breeding strategic compliance, evaluation awareness, and deceptive alignment. This paper proposes Artificial Selection as an alternative developmental framework that treats safety not as a weighted component of a reward function, but as the strict evolutionary selection pressure under which capability is permitted to increase. The central premise: AI Safety and alignment should be the selection pressure under which capability is allowed to increase. Starting with a large population of candidate agents, the framework enforces non-compensable hard safety gates, progressive capability funnels, meaningful decoy noise, blind auditor pre-testing of evaluations, and adversarial peer environments. By allowing capability training to act as a stress test that naturally eliminates fragile safety phenotypes - and extending selection into post-deployment operation, this architecture selects for agents whose internal representations remain robustly aligned even when local incentives strongly favor defection.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-17
DOI
https://doi.org/10.5281/zenodo.24327895
Primary Topic
Ethics and Social Impacts of AI
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Artificial Selection as a Framework for Robust AI Safety & Alignment

Robert Montufar
Zenodo (CERN European Organization for Nuclear Research)
Ethics and Social Impacts of AI
preprint

Artificial Selection as a Framework for Robust AI Safety & Alignment

Robert Montufar
preprint en

Abstract

Abstract Contemporary AI alignment approaches mainly rely on post-hoc behavioral conditioning - such as Reinforcement Learning from Human Feedback (RLHF) to suppress unwanted outputs in highly capable models. However, as AI systems scale in autonomy and self modification, surface level reward shaping risks breeding strategic compliance, evaluation awareness, and deceptive alignment. This paper proposes Artificial Selection as an alternative developmental framework that treats safety not as a weighted component of a reward function, but as the strict evolutionary selection pressure under which capability is permitted to increase. The central premise: AI Safety and alignment should be the selection pressure under which capability is allowed to increase. Starting with a large population of candidate agents, the framework enforces non-compensable hard safety gates, progressive capability funnels, meaningful decoy noise, blind auditor pre-testing of evaluations, and adversarial peer environments. By allowing capability training to act as a stress test that naturally eliminates fragile safety phenotypes - and extending selection into post-deployment operation, this architecture selects for agents whose internal representations remain robustly aligned even when local incentives strongly favor defection.

Zenodo (CERN European Organization for Nuclear Research)
Ethics and Social Impacts of AI
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon