Team 24 - VeriShield

Inspiration

With the overwhelming volume of online content today, detecting misinformation alone isn't enough. VeriShield not only identifies false claims but also prioritizes critical narratives that pose the greatest risk to organizations, ensuring key personnel are alerted promptly.

What it does

VeriShield’s retrieval engine, powered by a scheduled cron job, continuously scans media in real-time to capture emerging narratives. It evaluates threats from 'No Threat' to 'Critical', with clear explanations and credible sources. VeriShield also integrates deepfake detection for images and offers a tailored response generator, customizable to suit an organization's needs.

How we built it

We leveraged the T3 Stack for its type safety and streamlined development. For database interaction, we used Drizzle ORM, while Clerk ensured secure authentication. Tailwind CSS and ShadCN/UI provided a modern and responsive UI. Additionally, we employed Terraform to define and provision our infrastructure as code, enhancing scalability and consistency across environments.

Challenges, Learning and Accomplishments

Misaligned feature understanding underscored the need for mockups and wireframes to ensure clarity. Exploring new technologies such as AWS Lambda was challenging but rewarding, expanding our initial technical scope.

What's next for VeriShield

Future plans include integrating with social media for threat tracking and enhancing deepfake detection with advanced AI for improved accuracy in video, voice, and image analysis.