---
id: "concept-agi-automation-threshold"
type: "concept"
source_timestamps: ["¶3"]
tags: ["agi", "automation", "definitions"]
related: ["concept-service-as-software", "claim-agi-profit-reallocation", "quote-agi-definition", "entity-toby-stuart"]
definition: "AGI is defined pragmatically as the point at which the majority of tasks currently performed by humans on a computer become fully automated."
speakers: ["Toby E. Stuart"]
sources: ["futures"]
sourceVaultSlug: "hbr-seg-futures"
originDay: 2
articleStem: "hbr-nm-99-genai-end-incumbent-advantage"
sourceUrl: "https://hbr.org/2024/11/could-gen-ai-end-incumbent-firms-competitive-advantage"
sourceTitle: "Could Gen AI End Incumbent Firms’ Competitive Advantage?"
---
# AGI via Task Automation Threshold

[[entity-toby-e-stuart|Toby E. Stuart]] supplies a deliberately pragmatic, economic definition of Artificial General Intelligence. Rather than anchoring AGI to consciousness or philosophical human-equivalent reasoning, he holds that AGI *"has arrived in full force when the majority of tasks we now perform on a computer become automated"* (see [[quote-agi-definition]]). This is a **threshold definition**, and it is useful for business strategy precisely because it ties the arrival of AGI directly to the disruption of knowledge work and the subsequent uprooting of major social and economic systems. Once the threshold is crossed, the author argues, the software-driven disruption of the past two decades will *pale in comparison* to the ensuing changes.

This definition is the load-bearing premise for the rest of the argument: it is what makes the shift to [[concept-service-as-software|Service as Software]] inevitable and what drives the radical [[claim-agi-profit-reallocation|reallocation of profits]] across firms and industries.

**Enrichment / Validation.** The task-automation framing is *supported as a legitimate economic/strategic definition* but is *non-standard* relative to canonical AI-research definitions (which center on human-level generality, transfer, and novel problem-solving). Daron Acemoglu et al.'s *"Scenarios for the Transition to AGI"* (2024) formalizes an **automation index** and a threshold region above which additional automation begins to reduce wages and fundamentally changes macroeconomic dynamics — conceptually aligned with Stuart's framing. A canonical-AI counter-point: automating "the majority of tasks we perform on a computer" may *not* constitute true AGI if those tasks are narrow and lack broad generalization. Best labeled **economic/automation AGI**, distinct from full cognitive AGI.


## Related across articles
- [[concept-living-intelligence]]
- [[claim-compute-scaling-rate]]