← Back to Tutorials

3. Types of Artificial Intelligence

Getting Started

Based on Capabilities

Narrow AI (Weak AI)

Designed for a specific task. Examples include Siri, Alexa, chess engines, recommendation systems, and spam filters. All current AI systems are Narrow AI. They excel in their domain but cannot generalize beyond it.

General AI (Strong AI)

Hypothetical AI that matches or exceeds human intelligence across all cognitive tasks. It would understand, learn, and apply knowledge to any problem — just like a human. No General AI exists today.

Super AI (Artificial Superintelligence)

Theoretical AI that surpasses human intelligence in every domain — creativity, problem-solving, social skills, and scientific discovery. Purely speculative at this point, but a topic of active philosophical and safety discussion.

Based on Functionality

Reactive Machines

Pure reactive systems with no memory of past events. Example: Deep Blue (chess) — it analyzes the current board position but doesn't remember previous games.

Limited Memory

Can use past data to inform decisions. Self-driving cars observe recent speed and trajectory of nearby vehicles to make decisions. Most ML models fall here.

Theory of Mind

Hypothetical AI that understands human emotions, beliefs, and intentions. Would interact socially like another person. Not yet achieved.

Self-Aware

AI with consciousness and self-awareness. Purely theoretical and raises profound philosophical questions about machine rights and consciousness.

TypeMemorySelf-AwareExists Today?
ReactiveNoNoYes
Limited MemoryYesNoYes
Theory of MindYesNoNo
Self-AwareYesYesNo
Practice Task: Classify these AI systems by type (Narrow/General/Super) and functionality (Reactive/Limited Memory): ChatGPT, Roomba vacuum, TikTok recommendation algorithm, Google Search, a thermostat with a schedule. Explain your reasoning for each.