MCG CyberForce - Phishing Email Detector

Have you ever clicked an unknown link and then later on found out that your credit card has over a $5000 transaction? With MCG CyperForce, you can easily detect these phishing emails in no time.

Comment

Inspiration

  • With advancements in AI, phishing emails can become much more sophisticated. With simple web scraping and using ChatGPT, attackers can create a seamless email that you cannot resist the urge to click. Therefore, we have made the spam detection process much easier using the Chrome extension and web application - MCG CyperForce.

What it does

  1. Scan through all the links in an email.
  2. Using Google's Safe Browsing Lookup API to compare these links to Google's existing database of phishing emails.
  3. Output links that are identified as phishing URLs.

How we built it

For the Chrome Extension, we built it using JavaScript and Google's Safe Browsing Lookup API. For the web application, we use HTML, CSS, and JavaScript.

Challenges we ran into

Initially, we tried to train the ChatGPT Machine Learning model to identify spam and non-spam emails using different public datasets. Unfortunately, most of these datasets are rather limited and outdated. Not just that, spam emails nowadays, with the development of AI technology like Bard or ChatGPT, have been rapidly evolving in complexity, therefore it is impossible to detect them all precisely. Thus, we changed our approach to focus only on detecting whether links in a given email have been previously detected as malware. Then during the new approach, we ran into newer problems, the web scrapping data - URLs, takes all of the URLs in our Inbox rather than the email per se. We managed to work around it by developing a de-duping process. Last but not least, before we recorded our demo, our extension unexpectedly stopped, and we had to "calmly" debug, fortunately, we realized that Gmail itself was no longer recognizing the phishing URL in our test as a URL anymore.

Accomplishments that we're proud of

Despite our limitations in technical skills, we were able to finish our application with the help of many mentors as well as our own determination and passion. We had our own visions, and we exceeded our own expectations to achieve them

What we learned

Every member of our was able to extensively expand our knowledge in Machine Learning, Website Development, Cyber Security, and Application Development. We exposed ourselves to a wide variety of programming languages including Python, HTML, CSS, JavaScript, and API.

What's next for MCG CyberForce

We look forward to further developing our unit testing model as well as improve on the user’s interface and scalability of the program that we are implementing. We are also working towards improving our technical skills to better realize our ideas and visions. This can be our first Hackathon in TAMU, but it will most certainly not be our last!

Thank you very much, and Howdy!

Built With

Share this project:

Updates