← All ideas
Canon

Artificial General Intelligence (AGI)

framework · Computing Machinery and Intelligence (1950)

Confidence: High

AGI refers to a hypothetical type of AI that can understand, learn, and apply intelligence across a wide range of tasks at a human level or beyond. The concept originated from early AI research and remains a goal and subject of debate, with milestones like GPT-6 claiming to have achieved it.

Core Concepts

The Problem

No consensus exists on when or whether AGI will be achieved, leading to uncertainty about societal impacts.

The Claim

AGI will be achieved when an AI can perform any intellectual task that a human can.

Key Evidence

  • Turing's 1950 paper proposed a test for intelligence; milestones like GPT-4 and GPT-6 show rapidly improving AI capabilities.

Practical Implication

Achieving AGI would fundamentally change economy, labor, and ethics.

Nuance & Limits

Some argue current large language models already exhibit AGI-like behavior, while others say true AGI requires generalization and common sense.

Source Material

Citation Density

very high

Discuss Further

Open this concept in an AI assistant for deeper discussion, critique, or exploration.

Was this useful?