Inspiration
Phishing attacks are becoming more sophisticated, especially with AI-generated emails and malware campaigns like the Emotet Trojan. We realized that while hackers can spoof email addresses, they struggle to perfectly replicate someone’s unique writing style. That insight inspired Lion Lock AI: a smarter way to verify identity through language itself.
The logo was derived from the TCNJ mascot and the NJCCIC presentation sparked our interest in the cybersecurity area for this project, causing us to tie in the main focused idea into that.
What it does
- Lion Lock AI is a cybersecurity platform that compares the writing style of old emails with a new suspicious email to determine whether they were written by the same person. Using artificial intelligence and stylometric analysis, it examines tone, vocabulary, sentence structure, and linguistic patterns to detect impersonation. If inconsistencies are found, the system flags the email as potentially malicious, helping prevent phishing, scams, and malware infections.
How we built it
- We started by using GitHub Copilot extension on VS Code which created a baseline Python code for our website. We then debugged and added small tweaks to make it more polished. Then, we added to CSS and HTML to fix the interface to make it more user friendly. Lastly we integrated Gemini AI by using an API key.
Challenges we ran into
- One major challenge was ensuring accuracy with limited text samples. Writing style detection becomes harder with shorter emails. Another challenge was minimizing false positives, since natural tone shifts can occur depending on context or urgency. Balancing sensitivity with reliability required careful tuning of our model. After using VS Code, we also had many problems which we had to debug.
Accomplishments that we're proud of
- We are proud that we learned how to properly use Python in a coding project, as we are not fully fluent in it. We overcame many obstacles and this is our longest coding project yet. We also learned how to use an API key which will make projects much smoother in the future.
What we learned
- We learned how to use Python and also incorporate Gemini into our code. Plus we learned a lot about VS Code and how to use the platform with the GitHub Copilot extension.
What's next for Lion Lock AI
- Next, we plan to improve model accuracy with larger datasets and implement real-time browser or email integration. We aim to expand beyond this website and make it into an extension which can be applied to Gmail, Mail, etc. We also want to add additional security features such as password security checkers, data breach scanners, virus detectors, phishing detectors, website security checkers, etc.
NOTE: In our voiceover we said “app”, however Lion Lock is intended to be used as a website (please disregard that).
Log in or sign up for Devpost to join the conversation.