The most unusual AI Science projects students actually built

Think volcanoes and potato batteries are cool? Wait until you see what students are building with AI. From robots that react to emotions to smart tools that fix your math homework, the world of student science fairs is changing fast. In this post, we’ll explore the most unusual AI projects students have actually built, using powerful tools like Google’s Teachable Machine and other Google AI platforms. Ready to be amazed?

Why AI Science projects are the future of student innovation

What happens when kids are given access to powerful AI tools without needing to write a single line of code? Magic. Today’s students are no longer just building baking soda volcanoes, they’re building intelligent machines. Thanks to tools like Google’s Teachable Machine and other AI learning platforms for kids, science fairs are turning into showcases for creative, mind-blowing tech.

What makes a student AI project “Unusual”?

It’s not just about using AI, it’s about how students combine their imagination with technology to solve real problems or just have fun in unexpected ways. “Unusual” projects often involve mixing AI with art, nature, or even math, and they push the boundaries of what we expect from student work.

Meet the star: Google’s Teachable Machine

How students use Teachable Machine without coding

Google’s Teachable Machine is a web-based tool that lets students train machine learning models by simply feeding it examples, no code required. Want to teach your computer to recognize dance moves, animal sounds, or facial expressions? Just upload examples, train it, and go. It’s that easy.

You can try building your own AI model at Google’s Teachable Machine — no code required.

Real-world projects built with Teachable Machine

• A robot pet that reacts to human emotions
• A game controller that recognizes different hand gestures
• A voice-activated timer for visually impaired students

These projects aren’t just creative, they’re functional. Students are using this tool to explore the real-world impact of AI at a young age.

Other Google AI tools students are using creatively

Google AutoML: Can students train real AI models?

With Google AutoML, more advanced students can go beyond drag-and-drop tools and start building custom models. Some have used it to classify images of plant diseases, recognize objects for sorting systems, or even build simple recommendation engines.

Google Colab: A playground for young coders

Students who want to get their hands dirty with real code often turn to Google Colab. It’s like a notebook in the cloud where Python code meets machine learning. Many student coders use it for math-heavy AI experiments, like predicting equations or modeling real-world systems.

Bard AI and experiments with language-based projects

While Google Bard is often seen as a chatbot, students are starting to use it creatively, for language translation games, storytelling bots, and brainstorming AI-generated science fair concepts. It’s a playground for linguistic creativity.

Top 5 most unusual AI Projects students actually built

1. A plant that dances to human emotions

This project uses Teachable Machine to classify facial expressions like happy, neutral, or surprised. When the AI detects a happy face through the webcam, it triggers music and sends signals to servo motors connected to a fake “plant” (via Arduino or micro:bit) to sway or “dance.” It’s a fun way to explore emotion recognition and robotics. The project mixes biology, psychology, and AI in one charming idea.

2. Smart trash can that sorts waste using image AI

Built using Google AutoML, this bin uses a webcam to scan your trash and sort recyclables from landfill.

Students collect images of different types of waste (plastic, paper, etc.) and label them. These are uploaded to Google AutoML Vision to train a model. A Raspberry Pi or laptop with a webcam is used to scan items in real-time and trigger a servo motor to direct waste into the correct bin. A great hands-on sustainability project.

3. AI that detects mood from facial expressions

This project also uses Teachable Machine to recognize facial emotions via webcam. Once an emotion is classified, a preloaded playlist is triggered using simple JavaScript or Python. Students can connect this to Spotify’s API or just use local music files to create a dynamic, mood-reactive music player. No smile? It plays your favorite upbeat track.

4. Math pattern predictor using Google Colab

Using Google Colab and Python libraries like Keras or TensorFlow, students build a neural network that learns patterns from number sequences (like Fibonacci or arithmetic series). After training, the AI predicts the next number in a series faster than most humans could calculate. The dataset can be created manually or sourced from math problem repositories. A creative way to link AI and mathematics.

5. Robot arm controlled by hand gestures (No Code!)

Students train Teachable Machine to recognize different hand gestures via webcam. These gestures are mapped to specific commands sent to a microcontroller (like Arduino) connected to a DIY robot arm made from cardboard and motors. Tools like Pictoblox or Scratch extensions can help avoid coding altogether while controlling the hardware.

Bonus: Extra unusual projects worth exploring

Predicting stock market trends with Google Sheets + AI

This project uses Google Sheets to collect historical stock data (easily imported from finance APIs or Google Finance). Students can connect the sheet to Google Colab, where they run basic linear regression or LSTM models using TensorFlow. Visualization of predictions is done directly in Sheets or with Matplotlib. Great for combining math, economics, and machine learning.

AI that solves Geometry problems from drawings

Students can use Google AutoML Vision to train a model that classifies geometric shapes drawn on paper. Using a phone camera, they collect images labeled as “circle,” “square,” etc., and upload them to AutoML. Once trained, the AI can recognize the shape, and a linked app or script can calculate perimeter or area using predefined formulas. This combines computer vision with geometry. A great example of AI in math learning.

Training AI to recognize Algebra mistakes in homework

By inserting examples of incorrect vs correct algebraic equations into a model, this project helped solving math problems, a teacher’s dream tool.

Students can use Google Colab and build a simple classification model using Python (e.g., scikit-learn). By feeding the model expressions as text and labeling them as right or wrong, the AI learns to spot mistakes. With enough examples, the model can give feedback like “missing variable” or “sign error”, a great tool for peer tutoring or classroom use.

Language translator that learns from your voice

One ambitious student used Google Colab and voice datasets to build an AI that listens to spoken input and learns how to translate it across multiple languages.

To build this, students can collect voice samples using a microphone and label them in a CSV file. Then, using Google Colab and Python libraries like TensorFlow and SpeechRecognition, they can train a simple model that recognizes spoken input and maps it to phrases in another language. Over time, the model improves with more samples. It’s a fun way to explore voice AI and language learning.

How you can start your own wild AI Science fair project

You don’t need to be a coding wizard or math genius to build something jaw-dropping. Tools like Google’s Teachable Machine, Colab, and Bard are free and easy to use. Start small, train a model to recognize pets or voices, and then scale up. Who knows? Your idea might be the next project that ends up on this list.

As you’ve seen, today’s student innovators are thinking way outside the science fair box. With Google AI tools like Teachable Machine, AutoML, and Colab, building something brilliant doesn’t require advanced skills, just imagination. If these projects inspired you, it’s your turn to create something unforgettable. Explore the tools, test your ideas, and let your AI project take center stage!