An Artificial Intelligence URL Parser for Safer Web Browsing and Detection of Suspicious Links


Corresponding Authors:
James Jin1, Gayatri S2 and Yu Sun3, 1USA, 2University of California, USA, 3California State Polytechnic University, USA

Abstract

With more than seven billion people actively using the Internet, the number of cyber attacks has increased, and personal data breaches have become a concern among the general public. The COVID-19 pandemic has only increased the use of online platforms and services for work and leisure activities, which opens the door to more scams, viruses, and other cyber security breaches. Guided by SEO techniques and research regarding dangerous website and domain patterns, we have designed and implemented a visual system that tracks suspicious links on an active webpage and marks them in order to alert users to proceed with caution. Our AI utilizes linear regression to best detect trends in URL parsing, comparing them with registered unsafe links to see if they pose similar threats. The results reveal that AI isn’t entirely accurate since some trends are hard to decipher; however, it can reliably flag certain redirects and out-of-domain links that would otherwise remain hidden to users.

Keywords

Safe web browsing, Security application development, Phishing prevention.