Activities
Use any activity on its own, or run them in order as a one-hour lesson.
See Your Voice
A live picture of sound. Students whistle, hiss and whisper to discover what a microphone gives a computer.
Spectrogram Detective
Read sound pictures like a computer: use clues about hiss, stripes and gaps to match eight mystery pictures to their words.
Be the Recognizer
Students become the voice assistant, add noise and measure their word error rate, the same metric engineers use.
How Voice Assistants Hear You
A 16-slide visual tour: sound as a wiggle, the seven AI systems inside a voice assistant, and open problems that map to STEM careers.
Free tools to explore speech AI
Free tools for every stage of a voice assistant, with a privacy note for each, to test real systems against your students.
One-hour plan: an Hour of AI, speech edition
Fits Code.org’s Hour of AI in early December, or any science or computer science class.
- See Your Voice (10 min). Whistle, hiss and whisper. Big idea: a computer only gets a wiggle of air pressure.
- Spectrogram Detective (15 min). Read sound pictures with clues. Big idea: sounds leave patterns a machine can learn.
- Be the Recognizer (15 min). Be the voice assistant in quiet and noise. Big idea: noise and rare words cause errors, measured as word error rate.
- Human vs. machine (10 min). Read the same speaker cards to Windows Live Captions or phone voice typing, paste what it wrote into the online scorer, and compare error rates. Follow your school’s technology policy.
- Careers (5 min). Each of the seven systems inside a voice assistant is a STEM career (slide 7). Who will fix what voice assistants still get wrong?
For teachers
- Connects to AI4K12 Big Idea 1 (Perception) and Big Idea 4 (Natural Interaction).
- Printouts work in black and white.
- Students never need an account. See Your Voice runs entirely in the browser and records nothing.
- Free to print and share for classroom use.
About
Made by Deepak Babu Piskala, IEEE Senior Member and author of Building Speech AI, who worked on speech recognition for Amazon Alexa. First shared at the IEEE STEM Summit 2026 virtual exhibit booth. A personal volunteer project; views are the author’s own, not those of his employer or IEEE.
Questions or classroom stories? Get in touch.