ClarifAI

Inspiration

Our inspiration comes from the growing challenges users face in distinguishing authentic content from AI-generated media, fake news, and emotionally manipulative videos on platforms like TikTok. We observed that traditional fact-checking tools are often too heavy or disruptive, making them impractical for everyday use. Drawing on Apple’s Dynamic Island concept, we aimed to design a lightweight, seamless plugin that integrates directly into the viewing experience. The idea is to provide subtle but meaningful cues that guide users to pause, reflect, and, when necessary, access deeper contextual information.

What it does

Our plugin is designed to help users critically engage with digital content by detecting not only AI-generated media, but also signs of strong political bias, inflammatory language, and potential misinformation. Through subtle, real-time prompts and easy access to fact-checked sources, it empowers users to pause, reflect, and make more informed judgments, supporting healthier digital habits and protecting vulnerable groups from being misled.

How we built it

We followed the Double Diamond design process, starting with research and problem framing before moving into concept development and prototyping. Using Figma, we created multiple iterations of wireframes and interface designs, exploring how a Dynamic Island–style plugin could be seamlessly integrated into the app experience. Finally, we conducted user testing to evaluate usability, trust, and long-term value, gathering feedback that helped us refine features such as the alert page, visual prompts, and error-report button.

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for ClarifAI

We plan to expand the plugin’s functionality and reach. First, we aim to integrate it directly into popular platforms like TikTok and YouTube to maximize real-world impact. Second, we will refine the AI detection models to cover broader categories such as misinformation, political bias, and emotional manipulation, while reducing false positives. Third, we plan to introduce a community feedback system, allowing users to report errors and contribute to improving the accuracy of the plugin. In the long term, we envision building an educational mode that helps users, especially vulnerable groups such as the elderly, develop stronger media literacy and critical thinking skills.