Inspiration
Having spent 15 years in a small town facing severe cyberbullying, I've witnessed its devastating effects. Now, I'm participating in a machine learning competition to create tools that detect and prevent cyberbullying, ensuring no child suffers as I did.
What it does
How we built it
Frontend: Streamlit for a lightweight, interactive chat interface. AI model: Hugging Face toxicity detection to analyze text in real time. Data handling: Local JSON storage to save evidence securely. Gamification: A “shield” system that grows stronger as users block/report toxic messages, creating a sense of empowerment.
Challenges we ran into
Designing a user-friendly interface that feels safe and supportive for teens. Time constraints: building a demo that works across devices while keeping it simple enough to test.
Accomplishments that we're proud of
Built a working real-time chat prototype that highlights toxic messages. Designed a simple one-click system for saving evidence and blocking bullies. Created a concept that balances technology + empathy, empowering teenagers to feel protected online. Established a roadmap for scaling into multi-platform integration (WhatsApp, Telegram, etc.).
What we learned
How to combine AI models with simple UIs to solve real human problems. The importance of designing for emotional safety, not just technical accuracy. Hackathons are not just about code — they’re about storytelling, impact.
What's next for Cyber Shield Mission / "FriendShield"
Integrate with popular messaging platforms (WhatsApp, Telegram, Discord). Expand AI to detect bullying in images, voice, and memes. Build a teacher/parent dashboard with anonymized reports and early warning signs. Add multi-language support for broader accessibility. Partner with schools and NGOs to pilot the app in real communities.
Log in or sign up for Devpost to join the conversation.