Artificial Intelligence (AI) is the simulation of human intelligence by machines, particularly computer systems. It encompasses learning (acquiring knowledge), reasoning (using knowledge to draw conclusions), problem-solving, perception, and language understanding.
John McCarthy, who coined the term in 1956, defined AI as "the science and engineering of making intelligent machines."
Goals of AI
Reasoning — Draw logical conclusions from available information
Knowledge Representation — Store and organize information for inference
Planning — Set goals and create sequences of actions to achieve them
Learning — Improve performance based on data and experience
Natural Language Processing — Understand and generate human language
Perception — Interpret sensory data (vision, audio, touch)
Motion & Manipulation — Control robots and physical systems
Approaches to AI
Symbolic AI (Good Old-Fashioned AI) — Uses explicit symbols and rules for reasoning. Strong for logic and expert systems.
Connectionist AI (Neural Networks) — Inspired by the brain's neural structure. Excels at pattern recognition.
Evolutionary AI — Uses genetic algorithms and evolutionary principles for optimization.
Statistical AI — Relies on probability and statistics for learning from data (ML).
Hybrid Approaches — Combine multiple paradigms for complex problems.
AI vs Machine Learning vs Deep Learning
AI is the broad field. Machine Learning is a subset of AI where systems learn from data. Deep Learning is a subset of ML using multi-layered neural networks. Think of it as concentric circles: AI ⊃ ML ⊃ Deep Learning.
Why AI Now?
Three factors converged: massive data availability, powerful computing (GPUs/TPUs), and algorithmic breakthroughs (transformers, backpropagation, attention mechanisms). These have driven rapid advances in image recognition, natural language processing, and generative AI.
Practice Task: Identify 5 examples of AI you interact with daily (e.g., virtual assistants, recommendation systems, email spam filters). For each, classify which AI approach it likely uses (symbolic, connectionist, statistical, or hybrid). Write a one-paragraph definition of AI in your own words.