Click With Caution

homepage
result1
uploading a file
result2

Inspiration

We were inspired to make a tool to help people avoid falling victim to online scams. Initially, it was going to be marketed towards the elderly as a tech support bot. Eventually, we settled on the idea for an app to help everyone avoid being scammed.

What it does

Our website allows users to detect fraud. You may interact with the chat bot, upload screenshots and paste URLs into the chat. Our website uses AI to scan the provided documents and generate a scam report. The scam report tells you your probability of being scammed and an explanation of why it came to that conclusion.

How we built it

For the AI, we used Google's generative AI model Gemini. We embedded a chat box that connects to Gemini using a Python script. The front-end was developed in CSS and the back-end was developed in Python.

Challenges we ran into

At first, Gemini was unresponsive. Connecting the model proved to be the first major challenge. Creating a GitHub repository was the next big step. There were problems committing the contents of the project. After that, editing the UI was another hurdle.

Accomplishments that we're proud of

Since none of us had experience with AI, we're very proud to incorporate our first AI-driven tool. We are also proud to experience our first hackathon, as three out of our four members have never been to a hackathon before.

What we learned

We learned about working with large AI models, UI design, and web app development. Additionally, the experience working as a group with a tight deadline helped us learn about collaboration, setting realistic goals, and working under pressure. Finally, we learned how to deal with unexpected complications.

What's next for Click With Caution

Make the interface look simpler and easier to use. Adding more pre-select options, allowing users to specify if they want to scan for scams, fraud, phishing, viruses, or other online dangers. Future features could include the ability to enter website URLs, screenshots, and other files. Ultimately, our final vision for this project would be to have it embedded in web browsers and email services (such as Gmail). Licensing AI-based scam detection to companies so they can have it built in to their systems is the best model for a product like this, allowing anybody to browse the web safely.