Types of AI
Narrow vs general AI, and the capability levels in between.
- ✓The difference between Narrow AI and General AI (AGI)
- ✓Capability levels: reactive and limited-memory systems
- ✓Where today's systems actually sit
Explanation
AI is grouped by capability. Narrow (or weak) AI does one task well — spam filtering, translation, recommendations. Every system in use today is narrow AI, even very impressive ones.
General AI (AGI) would match a human across any task. It does not exist yet and remains a research goal.
Researchers also describe levels: Reactive machines respond only to the current input (like a chess engine), while Limited Memory systems use recent history (like a self-driving car tracking nearby cars). 'Theory of Mind' and self-aware AI are still theoretical.
A voice assistant feels general, but it is a bundle of narrow models — speech-to-text, intent detection, and text-to-speech — stitched together.
- • Thinking today's chatbots are 'general AI' — they are very capable narrow systems.
List three apps you used today and label each as narrow AI or not.
1. Which best describes today's AI systems?