Overview
Artificial Intelligence (AI) is the quest to build a brain. It is the science of making machines smart.
Core Idea
The core idea is Simulation. Can we simulate the cognitive functions of a human mind (learning, reasoning, problem-solving) in a machine?
Formal Definition
The study of “intelligent agents”: any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals.
Intuition
- The Calculator vs. The Poet: A calculator is better than you at math, but it’s not “intelligent” in the AI sense. AI aims for the messy, fuzzy stuff: recognizing a face, driving a car, writing a poem.
- Narrow vs. General AI:
- Narrow AI (ANI): Good at one thing (Chess, Spam filtering). We have this now.
- General AI (AGI): Good at everything (like a human). We don’t have this yet.
Examples
- Deep Blue: Beat Kasparov at Chess (1997). (Brute force search).
- AlphaGo: Beat Lee Sedol at Go (2016). (Deep Learning + Reinforcement Learning).
- ChatGPT: Generates human-like text. (Large Language Model).
Common Misconceptions
- Misconception: AI is conscious.
- Correction: Current AI is just math. It doesn’t “know” anything; it just predicts the next word or pixel.
- Misconception: It will kill us all (Terminator).
- Correction: The risk is real (Alignment Problem), but it’s more about “unintended consequences” than “evil robots.”
Related Concepts
- Machine Learning: The current dominant approach to AI.
- Turing Test: Can a machine fool a human?
Applications
- Healthcare: Diagnosing diseases from X-rays.
- Transportation: Self-driving cars.
- Creativity: AI art and music.
Criticism and Limitations
- Hallucination: AI models often confidently state falsehoods.
- Job Displacement: Automation of cognitive labor.
Further Reading
- Artificial Intelligence: A Modern Approach by Russell and Norvig
- Superintelligence by Nick Bostrom