Inspiration
In today’s digital world, misuse of technology — from unethical browsing to online fraud — often happens silently behind screens. We wanted to create a smarter, non-invasive way to understand human intent through natural cues. That’s how Whispy AI was born — inspired by the idea that our eyes can reveal more than words, serving as a window to digital ethics and awareness.
What it does
Whispy AI detects and analyzes a user’s pupil movements to identify suspicious or misuse-related behavior while interacting with digital systems. Instead of scanning full images or faces, it focuses purely on the eyes to protect privacy. It acts as a real-time guardian that can alert, restrict, or guide users when misuse is detected — ensuring safer and more responsible screen interaction.
How we built it
We used MediaPipe for accurate eye and pupil tracking and integrated it with Python for data analysis and behavior interpretation. The logic layer analyzes blink patterns, gaze direction, and focus levels to detect potential misuse. The prototype was built with OpenCV, Flask, and basic machine learning models trained on user interaction patterns.
Challenges we ran into
Achieving accurate detection under varying light conditions. Balancing privacy with real-time analysis — ensuring no sensitive images are stored. Designing an intuitive interface that doesn’t distract users. Training the model to differentiate between natural eye movement and actual misuse intent
Accomplishments that we're proud of
Successfully built a functional prototype that works without collecting facial data. Developed a privacy-first AI model that promotes ethical tech use. Presented the concept in cyber-security-focused hackathons and received strong positive feedback. Innovated a new perspective — analyzing human-computer interaction through eye behavior for security insights.
What we learned
We learned how small physiological signals, like pupil movement, can provide deep insights into digital behavior. We also understood the importance of building AI responsibly — focusing not just on detection accuracy but also user trust, ethics, and data privacy.
What's next for Whispy AI
We plan to expand Whispy AI with deeper behavioral analytics, adaptive learning for individual users, and integration into educational or workplace monitoring systems. Future versions will include an ethical dashboard and API for real-time misuse alerts, helping organizations build a culture of safe digital practices.
Log in or sign up for Devpost to join the conversation.