Two Big Buckets
Narrow AI (ANI)
AI designed and trained for a specific task. It can outperform humans in that narrow domain but does not have broad, general understanding.
- Examples: spam filters, facial recognition, recommendation engines
- Strengths: efficiency, accuracy in one domain
- Limits: can’t transfer skills to unrelated tasks
General AI (AGI)
A hypothetical AI with human-like general intelligence: it can learn any intellectual task a human can. Not available today; still a research goal.
- Examples: none (science fiction & active research)
- Potential: adaptable problem-solving across domains
- Challenges: safety, alignment, reliable reasoning
Behavior Types (Another View)
Reactive Machines
No memory; respond only to current input. Example: early chess programs that evaluate the present board.
Limited Memory
Use recent data/history to make better predictions. Most modern ML systems fall here.
Theory of Mind (Research)
Would understand emotions/intentions. Not yet achieved in practical systems.
Self-Aware (Speculative)
Would have consciousness/awareness. Purely hypothetical for now.
Where We Are Today
Almost all real systems today are Narrow AI with limited memory (they learn from data). They can be extremely capable within their domain (vision, speech, translation) but won’t seamlessly generalize to unrelated tasks without new training.