Inspiration

This project aims to protect individuals from Phishing attacks, specifically AI-generated attacks. A phishing attack is implemented by a malicious person acting as a reputable organization or person. These attacks usually implement some urgency, time commitment, or praying on non-tech-savvy persons. Misspellings, false emails, or fake logos can identify phishing attacks. With the ever-growing advancements of AI, our concern was the improvement in Phishing Attacks. Our theory is: "What if malicious individuals used AI text-generating tools to enhance their phishing attacks?" Thus eliminating the misspelling factor to identify a Phishing email.

What it does

📍identify if written by AI or human

📍identify logos

📍identify real and fake email language

📍identify real and fake email addresses

How we built it

Figma | HTML, CSS, JavaScript, PHP | Python | Google Cloud Vision API | Zyla API Hub AI Text Detection API | xamp

Challenges we ran into

API authentication issues, server-side API connectivity issue due to overloaded FIU wifi

Trouble connecting our API results to the front-end

Scope creep due to having many ideas we wanted to implement. Some of them we had to scrap due to technical/knowledge issues of implementation and implementation within the time constraint.

Sleep deprivation

Accomplishments that we're proud of

We built the website from scratch using HTML, CSS, JS, and PHP. We did not have to use a low-code/no code environment. Working with and implementing with technologies we had no experience before, specifically implementing the Google Cloud Vision API and using the ZYLA API Hub AI Text Detection API. The work flow of our project is almost functional.

we is alive. yay.

What we learned

A balanced team and concise communication make or break a project. If we did not have people sticking to the project charter we would have lost the focus on the core concept. On the other side if we did not have people pushing new ideas we would have been stuck trying to implement the ML models, which were to time intensive to implement, instead of relying on APIs that enabled us to have an almost finished mvp. we is good team. woohoo!

What's next for Themis

bring home da bread... $$$$$

Share this project:

Updates