Background

Phishing is a common and pervasive form of cyber-attack. A fraudulent message is often sent to target an individual or organization to obtain sensitive information or to deploy malicious software, such as trojans and ransomware. The level of sophistication of phishing attacks has significantly increased over the years. This sophistication has led to a vast number of individuals and organizations being affected by these attacks which often go undetected by current cybersecurity solutions. The impacts of phishing have consequences ranging from financial and reputational. Therefore, it is paramount to detect phishing emails to mitigate the spread of cyber-attacks in email and messaging platforms.

Technology Overview

The proposed patented technology is a software solution for automatic classification of electronic messages into multiple trust-based. The trust‑based folders to protect users of email and messaging platforms from phishing attacks and related cybercrime. It uses machine learning approaches to detect phishing emails by stylometric authorship verification. As the message leaves the email outbox, the technology verifies whether the message was drafted and sent by the email account owner and not by an impostor. In addition, as the message arrives in the email inbox, the message is analyzed for content, route, origin of message, attached file attributes, and uniform resource locators (URLs) embedded in the message. A risk score is computed and used to classify the incoming message into one of the folders with associated risk level.

Benefits

• Improves detection of phishing attacks
• Works well even with short messages
• Augments existing technology solutions to enhance cybersecurity
• Integrates well with existing email platforms (e.g. Gmail and Outlook)

Applications

• Email and messaging security

Opportunity

• Strategic business partnership with Email Veritas (https://www.emailveritas.com/), a British Columbia-based cybersecurity company that develops anti-phishing technologies