The best way to understand how AI works is to think about how you learn. AI systems learn in a similar way: by being given lots of information (data) and then being told what to do with it.
AI learns through a process called Machine Learning, which has three main parts:
1. Data
AI models learn from data. This data can be anything—pictures, text, sounds, or numbers. The more high-quality data an AI model has, the better it can learn. It’s like a chef learning to cook by tasting thousands of different ingredients and recipes.
2. Training
After collecting the data, the AI model is “trained” on it. This is a lot like a student studying for a test. The AI is given the data and an objective (for example, “find the difference between a dog and a cat”). It runs through the data over and over again, making adjustments and getting better with each try until it can accurately complete the task.
3. Inference
Once the AI model is trained, it’s ready to go! This is the moment it uses what it learned to do new things. When you take a picture with your phone and the AI recognizes a friend’s face, that’s inference. The AI is using what it learned during training to make a real-time prediction or decision.
LET’S REFLECT…
- Think about a game or app you use. How does its use of AI relate to the “Data,” “Training,” and “Inference” process?
- Now that you know how AI learns, how do you think a person’s digital footprint (what they post and share online) could be used to train an AI model?
Try out this Machine Learning Tool
Ready to try training your own AI? Follow the link below to play with a simple machine learning model.
- Teachable Machine lets you train a computer to recognize images, sounds, and poses with just a few clicks. You can give it your own data to train it and then see how it works!