1.5 Applications of AI (BT104CO)
1. Major Domains of Application
AI has transitioned from lab experiments to ubiquitous technology, powering systems that solve real-world problems through rational action.
A. Robotic Vehicles
The most visible application of the Rational Agent model. Involves high-speed sensor fusion and real-time path planning.
- Examples: Self-driving cars (Tesla, Waymo), Mars Rovers.
B. NLP & Speech
Interacting with machines using human language by understanding syntax, semantics, and context.
- Examples: Virtual assistants (Siri, Alexa), LLMs (ChatGPT).
C. Games & Planning
Searching massive state spaces and using Heuristics to evaluate optimal board positions.
- Examples: Deep Blue (Chess), AlphaGo (Go).
D. Medicine
AI acting as an expert assistant for pattern recognition and high-precision classification in diagnosis.
- Examples: Tumor detection (MRIs), AlphaFold (Proteins).
E. Finance & E-commerce
Using probabilistic reasoning and anomaly detection to predict human behavior and market trends.
- Examples: Fraud detection, Recommendation engines.
2. Classification by Function
In the context of exams, applications can be classified by the type of cognitive task they perform:
| Category | Description | Example |
|---|---|---|
| Perception | Interpreting binary sensory data. | Facial recognition, OCR. |
| Reasoning | Drawing logical conclusions. | Legal research, diagnostic systems. |
| Planning | Sequencing actions to reach a goal. | Supply chain optimization. |
| Learning | Improving based on experience. | Spam filters, personalized ads. |
3. The "AI Effect" (A Key Concept)
The Paradox of Success
Russell and Norvig note that "as soon as an AI technique becomes common enough, it is no longer called AI." This is known as the AI Effect. Optical Character Recognition (OCR) was once considered miraculous AI; today it is standard computation. "AI" often defines whatever hasn't been done yet.