CSR Home › AI for Beginners › Day 2
One step at a time. Finish the earlier days first.
Day 2 of 7
How Does AI Work?
Today you can:
- Explain data → learning → prediction
- Understand why more good examples help
- See where bias can come from
Every AI follows three steps.
- Datacollect examples
- Learningfind the pattern
- Predictionguess something new
One-sided examples give one-sided guesses. This is called bias.
Key words
Data: the examples the AI learns from
Training: the learning step
Model: what training produces
Seen in India
A spam filter learns from emails people marked as spam. It then guesses whether your new email is spam.
A crop-advice app that learned from photos of healthy and diseased leaves may guess a disease from a new photo. If it never saw your local crop, it may be wrong.
Guided activity: Which step is it?
For each sentence, choose the step of AI it describes.
Gathering examples is the data step.
Improving from mistakes on the examples is learning.
A guess on something new is a prediction.
That is a question about the quality of the data.
Again, a guess about something new.
Quick quiz
Back to CSR HomeSupport Children’s EducationWomen Career Initiative