GmailPhishGuardian AI

An AI-powered Chrome extension that safeguards Gmail by detecting phishing threats and securing your inbox.

Comment

Inspiration

The rise in phishing attacks has made email security a top priority for individuals and businesses alike. I was inspired to create GmailPhishGuardian AI because of my background studying Cybersecurity and because I know the importance of protecting against this #1 security threat, which is the users. To empower users with better tools with real-time protection against phishing threats right within their Gmail inbox and by leveraging AI and integrating seamlessly into Google Chrome, this project aims to provide users with a simple yet powerful solution to stay secure online.

What it Does

GmailPhishGuardian AI is a Chrome extension that integrates with Gmail to detect and analyze phishing emails in real-time. Here’s what it does:

  • Scans incoming emails for phishing red flags using AI-driven analysis.
  • Highlights potential threats and provides a phishing likelihood score as warnings to the users.
  • Helps users protect their data and avoid falling victim to scams with intelligent recommendations.

How We Built It

We built GmailPhishGuardian AI using:

  • Google Chrome Extensions API for seamless browser integration.
  • Google’s built-in AI Prompt API to detect phishing patterns, evaluate sender behavior, and analyze content for signs of deception.
  • Not only analyzes the general email content but also factors like sender, subject, body, and timestamp.
  • JavaScript and HTML/CSS for the user interface, ensuring ease of use and a smooth experience.

Challenges We Ran Into

  • Ensuring high accuracy of phishing detection while minimizing false positives took lots of hours of testing by re-designing the prompt.
  • Designing an intuitive interface to simplify phishing insights for users (Phishing Score and Analysis).
  • Testing and fine-tuning the AI to handle diverse email types and formats.

Accomplishments That We’re Proud Of

  • Successfully building an AI-powered Chrome extension that integrates directly with Gmail. I know it could be improved, but it was a challenge doing this the first time.
  • Creating a user-friendly tool that makes advanced phishing detection accessible to everyone.
  • Ensuring robust detection capabilities while maintaining user data privacy and security, as no data is being sent to the server.
  • Overcoming technical hurdles to deliver a polished and effective solution, especially since this was my first time working on something of this level.

What We Learned

  • How to effectively utilize Google APIs and Gemini Nano built-in AI (Prompt) and integrate them with Chrome extensions.
  • How to balance performance, accuracy, and usability in an AI-powered application. (Personal tip that could have saved me hours: Long Prompt = Long Time Waiting for Response.)

What’s Next for GmailPhishGuardian AI

  • Offer insights into suspicious links and improve sender authenticity.
  • Utilize Gmail APIs to analyze email content instead of the DOM and extract relevant metadata. I believe this will integrate better with the extension and is a secure method of extracting the data since it’s Google-driven.
  • Expand detection capabilities to handle more sophisticated phishing attacks by adding more complex parameters.
  • Add support for multi-language phishing detection to serve a global audience (once available or using the Gemini Translate API for this).
  • Enable real-time reporting and sharing of phishing threats among all users, making it easier to not only rely on AI to do the job but also on previous encounters.
  • Build additional features, such as email categorization and advanced threat analytics.
  • Expand my cybersecurity knowledge to refine the AI logic.
  • Launch the extension publicly on the Chrome Web Store to help more users stay safe online once this experimental stage is done.

Built With

Share this project:

Updates