Inspiration
The rise of AI tools like ChatGPT and Google Gemini has transformed learning—but also made it easy to bypass real understanding. We saw students submitting AI-generated work with little effort, while institutions lacked practical ways to ensure academic integrity without blocking learning. HumanFirst was inspired by one core question: How can we allow AI-assisted research but still ensure human learning?
What it does
HumanFirst is an intelligent academic integrity platform with two modes: Assignment Mode Allows open research Detects AI platform usage in real-time Warns students instead of blocking them Logs activity and generates risk scores Exam / Focus Mode Locks the device into a secure environment Prevents access to unauthorized apps and websites Ensures a fully controlled exam experience It balances freedom + accountability instead of strict restriction.
How we built it
Desktop application using Electron / .NET hybrid architecture Monitoring agent for system activity and domain tracking Backend powered by edge functions and database (e.g., Supabase) Policy-based control system for institutions Detection engine for AI platforms We are now transitioning toward a controlled in-app browsing environment using WebView2 for reliable tracking.
Challenges we ran into
Detecting AI usage in external browsers (due to HTTPS and OS limitations) Lack of reliable domain visibility without full environment control Balancing privacy with monitoring Avoiding intrusive or restrictive user experience Building a system that works in real-world academic environments
Accomplishments that we're proud of
Built a working prototype with dual-mode system Designed a non-restrictive AI monitoring approach Identified and solved core architectural flaws early Created a scalable foundation for real deployment Reframed the problem from “blocking AI” to “ensuring learning”
What we learned
Monitoring without control is unreliable Product design matters more than just technical implementation Users resist restrictive systems Real innovation comes from balancing opposing forces (freedom vs control) Early testing exposes fundamental flaws faster than theory
What's next for HumanFirst
Implement full in-app research environment using WebView2 Improve AI detection accuracy and expand domain database Build advanced analytics and behavior insights for institutions Add plagiarism + AI hybrid detection Pilot testing in schools and universities Scale into a complete academic integrity ecosystem
Log in or sign up for Devpost to join the conversation.