CMore

Plugin on an Instagram post- Prototype

Inspiration

With the rise of AI-generated content and misinformation, people are more vulnerable to scams than ever before. This is especially concerning for those with low social media literacy, such as pre-teens and elderly individuals. Our goal is to highlight original content and ensure it receives the credibility it deserves

What it does

An extension that identifies whether text is false or not and determines whether an image is AI-generated or altered
Features an icon with colour indicators to show if a post is true/real, uncertain, or false/fake
Results have explanations for why content may be false or why an image appears AI-generated
Results also include links to sources to help identify misinformation
Users can upvote/downvote the results to verify the model’s accuracy

How we built it

Prototype built on Figma!

Challenges we ran into

Detecting AI-generated text is more complex than it seems, as it requires a deep understanding of common LLMs. Given this challenge, we decided to shift our focus to AI-generated images instead.
No model can be 100% accurate, and false positives or negatives could lead to misinformation. To mitigate this, our plugin prompts users to verify the model's classifications, incorporating human input to enhance reliability.
Initially, we planned to provide an accuracy score for posts and content. However, to avoid misleading users with potential false positives or negatives, we opted for a color-coded system to indicate accuracy levels instead