Systems for Malicious Website Detection Using Machine Learning

ChangeBridge: Patent Grants - Networking (H04L)

Published March 31st, 2026

Detected March 31st, 2026

Summary

The USPTO granted Patent US12592967B2 to BTblock for systems that detect and prevent navigation to malicious websites using machine learning. The patent covers technology that extracts URLs, compares them against domain name databases of safe and unsafe sites, and uses machine learning models trained on updated databases to enhance detection capabilities. Invented by Thomas Olofsson, the patent contains 20 claims and was filed April 19, 2024 under application number 18641082.

View original document View source feed page

What changed

USPTO granted Patent US12592967B2 to BTblock for a machine learning system that detects malicious websites. The system extracts URLs, compares them against domain name databases containing data on safe and unsafe websites, and uses machine learning models to identify threats. When an unsafe website is detected, the system updates its databases and retrains the models to improve future detection. The patent includes 20 claims under CPC classification H04L 63/1483.

This is a patent grant notification, not a regulatory requirement. No compliance actions are required from other entities. Technology companies developing cybersecurity solutions may wish to review this patent to understand existing intellectual property in the malicious website detection space. The patent does not impose obligations or deadlines on third parties.

Source document (simplified)

← USPTO Patent Grants

Systems for malicious website detection using machine learning

Grant US12592967B2 Kind: B2 Mar 31, 2026

Assignee

BTblock

Inventors

Thomas Olofsson

Abstract

Systems are disclosed herein for detecting and preventing navigation of a user computer to malicious websites. The systems can detect malicious websites using machine learning models and domain name databases containing data on safe and unsafe websites. The systems can extract a uniform resource locator (URL) associated with a webpage and compare the URL with the domain name databases to determine if the URL is safe or unsafe. When the system detects an unsafe website, the system can update the domain name databases to include data on the new unsafe website. The machine learning model may be trained using the updated domain name databases to enhance the detection of malicious websites and prevent users from navigating to such websites.