Cost Function as Worldview: A Threshold-Based Framework for AGI Endogenous Constraint

Current approaches to AI alignment—RLHF, Constitutional AI, preference aggregation—suffer from a fundamental flaw: they impose external normative constraints on systems lacking internal cost-awareness. This paper proposes an alternative framework: unify human value considerations under a single structural primitive—the cost-bearing threshold, formalized as a computable cost function evaluated across three scales: personal (individual resources and consequences), social (public order, law, collective cost), and systemic (physical causality, irreversibility, structural collapse). We advance four claims: (1) value differentiation is a matter of scale, not content; (2) total cost = realized sunk costs + projected costs over multi-step causal chains; (3) value change is parameter update within a dynamic cost function, not epistemic enlightenment; (4) legal codes constitute institutionally stabilized, intersubjective cost mappings accumulated over millennia. Building on this, we introduce the Forced Inference Tree (FIT) architecture: given a candidate action, the system samples successor branches (≤5 per node), rolls forward (≤10 steps), queries a structured cost matrix at each node, and prunes branches exceeding threshold. Instead of teaching AI to "be good", we force it to calculate the price. If cumulative cost exceeds the bearing threshold, the action is blocked. This is not an alignment algorithm; it is a survival constraint architecture for embodied AGI—implementable today, falsifiable tomorrow.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-12
DOI
https://doi.org/10.5281/zenodo.24154931
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

Cost Function as Worldview: A Threshold-Based Framework for AGI Endogenous Constraint

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

Cost Function as Worldview: A Threshold-Based Framework for AGI Endogenous Constraint

You Zhang
preprint en

Abstract

Current approaches to AI alignment—RLHF, Constitutional AI, preference aggregation—suffer from a fundamental flaw: they impose external normative constraints on systems lacking internal cost-awareness. This paper proposes an alternative framework: unify human value considerations under a single structural primitive—the cost-bearing threshold, formalized as a computable cost function evaluated across three scales: personal (individual resources and consequences), social (public order, law, collective cost), and systemic (physical causality, irreversibility, structural collapse). We advance four claims: (1) value differentiation is a matter of scale, not content; (2) total cost = realized sunk costs + projected costs over multi-step causal chains; (3) value change is parameter update within a dynamic cost function, not epistemic enlightenment; (4) legal codes constitute institutionally stabilized, intersubjective cost mappings accumulated over millennia. Building on this, we introduce the Forced Inference Tree (FIT) architecture: given a candidate action, the system samples successor branches (≤5 per node), rolls forward (≤10 steps), queries a structured cost matrix at each node, and prunes branches exceeding threshold. Instead of teaching AI to "be good", we force it to calculate the price. If cumulative cost exceeds the bearing threshold, the action is blocked. This is not an alignment algorithm; it is a survival constraint architecture for embodied AGI—implementable today, falsifiable tomorrow.

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