---
id: "concept-augmentation-score"
type: "concept"
source_timestamps: ["¶9"]
tags: ["methodology", "labor-metrics", "task-analysis"]
related: ["framework-task-categorization-scoring"]
definition: "A metric calculating an occupation's potential for AI enhancement, based on the ratio of AI-exposed tasks to unexposed, human-dependent tasks."
sources: ["reskilling"]
sourceVaultSlug: "hbr-seg-reskilling"
originDay: 10
articleStem: "hbr-edu-35-ai-changing-labor-market"
sourceUrl: "https://hbr.org/2026/03/research-how-ai-is-changing-the-labor-market"
sourceTitle: "Research: How AI Is Changing the Labor Market"
---
# Augmentation Score

**Definition:** A metric calculating an occupation's potential for AI enhancement, based on the ratio of AI-exposed tasks to unexposed, human-dependent tasks.

The augmentation score is a metric developed by the research team to quantify an occupation's potential to be enhanced by generative AI. It is constructed by analyzing the specific **tasks** that make up an occupation and calculating the share of **'exposed'** tasks (those that can be automated or assisted by AI) versus **'unexposed'** tasks (those requiring strictly human involvement). This score is what differentiates jobs likely to be displaced ([[concept-ai-automation-displacement]]) from those that will evolve into human-AI collaborative roles ([[concept-ai-augmentation-complementarity]]).

The score is the output of the [[framework-task-categorization-scoring|task categorization methodology]] and presupposes the [[prereq-task-based-labor-model|task-based model of labor]] — that jobs are bundles of separable tasks.

**Enrichment / confidence note:** The working paper ([[entity-displacement-or-complementarity-paper]]) builds occupation-level exposure/augmentation metrics from the exposed-vs-unexposed task share, conceptually equivalent to this score. Anthropic ([[evidence-anthropic-labor-study]]) and the World Bank ([[evidence-world-bank-labor-demand]]) construct methodologically similar AI-exposure indices based on task feasibility with LLMs, validating the metric family.