Classes 8–10
Jugnupur Tales, Volume 2
How the City Learned About AI
Back home in Bengaluru, thirteen-year-old Bunny, his friend Riya and the whole Silverline Heights society run into AI everywhere: in a cricket-match vote, a foggy trek, a topper who only memorised, a chatbot's training, a sneaky forwarded message, a fooled pothole app, an endless video feed, a thirsty data centre and a school hackathon. Ten chapters on how neural networks really learn, how AI can be tricked, and how to build it responsibly.
Chapters
10 chapters
- The Society Committee VoteA neural network is many small weighted votes
- Fog on Nandi HillsA model learns by taking small steps downhill on its errors
- The Topper Who Only MemorisedA model must work on new examples, not just the ones it learned from
- A Map Made of WordsAI turns words into numbers, so that words with similar meanings sit close together · Getting published soon
- How a Chatbot Goes to SchoolA chatbot learns in stages, first knowledge, then manners, judgment and safety · Getting published soon
- The Forwarded Message With a Secret OrderAn AI assistant can be tricked by instructions hidden inside the things it reads · Getting published soon
- The Pothole App That Was FooledIf someone sneaks bad examples into an AI’s training data, the AI learns the wrong lesson · Getting published soon
- Just One More VideoRecommendation AI chooses what you see next, and it is trying to keep you watching · Getting published soon
- The Tanker and the Data CentreAI runs on real electricity and real water, so we should use it wisely · Getting published soon
- The School HackathonBuilding good AI means testing it in the real world, knowing when it’s unsure, and working with the people who know the job · Getting published soon
Your progress is saved on this device only. No sign-up needed. The 5-minute read and quiz for every chapter are free on Learn.