RP SOI students gamify deepfake detection to make AI learning accessible

"The goal of our FYP, Fool the AI Detector, was to make AI more accessible and less intimidating. We wanted to encourage experimentation and creative thinking in a way that feels more like play than practice," says Julian Wong on behalf of his team.

Comprising RP SOI students Julian, Sherval Ng, John Miyahara and Gideon Lim, the group developed a deepfake detection model that analyses user-generated images to determine whether they were created by AI. Beyond detection, the platform also teaches users how to spot manipulated images and shares tips on prompt engineering.

"After getting feedback from our classmates and our lecturer, Mr Chua Swee Kwang, we decided to gamify the model to make it more engaging and fun," Julian explains. "We introduced an interactive web game component where people can play and learn together, instead of going through a dry tutorial."

Recently, the team showcased their project at the RP Open House 2026, where they had the opportunity to introduce their work to secondary school students and visitors. Through live demonstrations and hands-on play, they shared how AI can be both powerful and approachable, sparking conversations about creativity and innovation in tech.

Designed for a wide audience, from students to AI enthusiasts, the team hopes the platform will help users sharpen their AI image generation skills through friendly, competitive challenges. Users receive real-time feedback and grades generated by AI models, allowing them to refine their prompts iteratively while building both creative confidence and technical skills.

Excited about creating innovative AI solutions? Discover how the School of Infocomm empowers you to turn your ideas into reality.