---
title: "What Is AGI? Meaning, Definition and Levels"
description: "What is AGI? Artificial general intelligence in plain words, how it differs from narrow AI, the five levels researchers propose and what AI does today."
url: https://nexumautomations.com/glossary/artificial-general-intelligence
updated: 2026-11-10
publisher: Nexum Automations, an AI consulting firm
---

> Nexum Automations is an AI consulting firm that builds what it recommends. It finds where AI pays off in a company, builds it on the tools the team already uses and stays until it is live and working.

Glossary

# What is AGI? Explained

A plain definition of artificial general intelligence, why it is still a concept, how it differs from the AI companies use today and how to judge AI by what it does now.

Last reviewed November 10, 2026

Definition

Artificial general intelligence (AGI)

What is AGI? Artificial general intelligence is a hypothetical kind of AI that could learn and do nearly any intellectual task a person can, across many fields, instead of being built for one job. Sources describe it as theoretical and not yet realized.

**In short:** Hypothetical AI that matches people across fields.

Our data

## What buyers see in this search

The first page of Google for "what is agi", counted by type of result.

| | Results |
| ---------------------------------------------------- | ------------------------------- |
| Articles by companies that sell a product or service | 4 of 9 |
| Forums and Q&A threads | 3 of 9 |
| Independent publishers and explainers | 2 of 9 |
| Titles that answer this exact search | 9 of 9 |
| Google features on the page | An AI Overview, People also ask |

The 9 organic results on the first page of Google, United States, desktop, measured with DataForSEO on October 29, 2026\. Types only: no company or product is named.

Side by side

## How the sources define AGI

Two sources, two ways to set the bar.

| | How it defines AGI | The bar it sets | Status as stated |
| ----------------------------- | ---------------------------------------------------- | ------------------------------------- | ---------------------------- |
| Rutgers AI Ethics Lab | Theoretical AI doing a wide range of cognitive tasks | At or beyond human-level intelligence | Not yet realized |
| Morris et al., DeepMind, 2023 | Systems classified by depth and generality | Percentiles of skilled adults | Only the first level reached |

## What AGI stands for and what it means

* Artificial
Built by people, not born.
* General
Works across many fields, not one.
* Intelligence
Learns and applies knowledge.

AGI stands for artificial general intelligence: a system that is general, not limited to one task.

The AGI meaning is easy to hold on to. Today's AI is built to do one thing well. AGI would carry what it learns from one field into another, with a person's flexibility.

AGI is also the tax term adjusted gross income. This page covers only the AI meaning.

## Why AGI is a concept, not a product

What the sources say

* Hypothetical or theoretical
* Spans many fields
* Would learn with some autonomy

What nobody can point to

* A product you can buy
* An agreed single definition
* A system the sources call AGI

The AGI definition depends on who writes it.

Rutgers' AI Ethics Lab calls it a theoretical, not yet realized form of AI that can handle a wide range of cognitive tasks at or beyond human level.

The same entry says there are differing opinions on whether it is feasible and when it might arrive.

Rutgers also uses AGI and superintelligence together, while other sources keep them apart. That is why definitions vary.

Researchers at Google DeepMind (Morris et al., 2023) take a different route.

They classify systems on the path to AGI by depth, meaning performance, and breadth, meaning generality. So no system today reaches the level most of these definitions mean.

## How AGI differs from close terms

Narrow AI

* Built for specific tasks
* Used by companies today
* Chatbots and assistants

AGI

* Works across many fields
* Described as hypothetical
* Learns like a person

Sources split AI by how wide it reaches. The AI companies use now, including generative AI and agentic AI, is narrow: strong at specific tasks.

AGI is the general step beyond that, and superintelligence is the step past people. For the base term, see [artificial intelligence](https://nexumautomations.com/glossary/artificial-intelligence).

## The five levels of AGI

1. Emerging
Equal to or somewhat better than an unskilled human.
2. Competent
At least the 50th percentile of skilled adults.
3. Expert
At least the 90th percentile of skilled adults.
4. Exceptional
At least the 99th percentile of skilled adults.
5. Superhuman
Outperforms all humans.

Morris et al. (Google DeepMind, 2023) in their paper Levels of AGI sort AI by how it compares with skilled adults.

Competent means at least the 50th percentile, Expert the 90th and Exceptional the 99th. Superhuman outperforms all humans. Only the first level has been reached, and the authors place the general chatbots of 2023 there.

The authors also write that their Competent level is probably the best catch-all for many existing AGI definitions.

## What AI can do today

1. Pick a task
Start with one repeatable task.
2. Test the AI
Check what it does well today.
3. Keep judgment
People decide the hard cases.
4. Measure
Compare the result with the old way.

For a business, the useful question is what AI does well now.

It is strong at specific tasks, so the practical way to look is task by task, with people keeping the judgment. [Our method](https://nexumautomations.com/roles/method) does exactly that, using task-level evidence.

Two ideas help here: [AI augmentation](https://nexumautomations.com/glossary/ai-augmentation), where AI extends what people do, and [AI native](https://nexumautomations.com/glossary/ai-native), where a workflow is built around AI from the start.

Our [AI consulting](https://nexumautomations.com/services/ai-consulting) team finds the places where AI pays off in your company, builds on the tools you already have and stays until the system is live.

FAQ

## AGI questions

* ### What does AGI stand for?
The letters mean artificial general intelligence here. In finance the same letters can mean adjusted gross income, a different thing this page does not cover.
* ### What is the definition of artificial general intelligence (AGI)?
Rutgers' AI Ethics Lab calls it a theoretical, not yet realized form of AI that can perform a wide range of cognitive tasks at or beyond human level. Unlike narrow AI, it would apply knowledge across different fields.
* ### What is AGI in AI compared with narrow AI?
Narrow AI is built for specific tasks, which describes the AI companies use today. AGI would work across many fields with its own autonomy. See [AI vs machine learning](https://nexumautomations.com/compare/ai-vs-machine-learning) for another common split.
* ### What are some examples of general AI?
None of the sources names one that exists today. Morris et al. rate only the first level as reached, and the higher levels as not yet achieved.
* ### Is AGI the same as superintelligence?
Sources differ. Rutgers uses the terms together, while the DeepMind paper treats superintelligence as the top level, above Expert and Exceptional, where a system outperforms 100% of humans.

---

Source: https://nexumautomations.com/glossary/artificial-general-intelligence

Free AI diagnosis (30-minute call): https://nexumautomations.com/#book - Services: https://nexumautomations.com/services - Guides: https://nexumautomations.com/guides
