Types of AI (Narrow, General, Super) — Tutorial

Narrow AI — the only kind that exists today

Narrow AI (also called Weak AI) is designed for one specific task. A chess engine beats world champions but cannot write an email. GPT-4 generates fluent text but cannot play chess at the same level. Image classifiers identify tumors but cannot drive a car.

Every AI system in production today — recommendation engines, fraud detectors, LLMs, image generators, self-driving prototypes — is Narrow AI. Each is a specialist, not a generalist.

Artificial General Intelligence (AGI)

AGI describes a hypothetical system that matches human cognitive ability across all domains — it could switch from writing code to diagnosing disease to learning a new language with the same facility a human can.

No AGI exists today. Definitions vary (some researchers argue frontier LLMs show early AGI precursors; most disagree). The AI safety field treats AGI as the central long-term risk to plan for, regardless of the timeline uncertainty.

Superintelligence and the media hype gap

Superintelligence would surpass human capability across all domains by a large margin. It is purely theoretical and no credible timeline for its arrival is established.

Most day-to-day "AI" headlines describe Narrow AI advances. Reading those stories with the Narrow/General/Super framework prevents confusion: a model that aces the bar exam is still Narrow AI — it was trained on legal text and optimised for that task.

Continue your research

Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Types of AI (Narrow, General, Super) — Tutorial.

• Browse models
• Review providers
• Inspect benchmarks
• Explore workflows
• Learn AI concepts
• Read decision guides
• Review trust and methodology
• Start a decision path