Machine learning has gone from rags to riches over the last decade: once dismissed as a computer scientist’s pipe dream, we now live in the time of the ‘Great A.I. Awakening’ where Google Brain, neural networks and the University of Toronto’s own Vector Institute are no longer niche references for computer scientists. Machine learning has become part of the mainstream cultural lexicon.
If you have access to a computer, machine learning has probably sparked your imagination. While we certainly aren’t living in an age of advanced artificial lifeforms and dystopic friendships with computers, machine learning’s potential applications excitingly include improved medical diagnoses and self-driving cars.
But what does machine learning mean for information security? It can mean a lot. Cyber security experts continue to identify new ways of using machine learning to better detect and prevent potential risks and threats to your information. Here are three:
- Machine learning can be used to detect phishing attempts. The social engineering element of phishing emails makes it tricky for computers to detect phishing emails: the most dangerous phishing attempts succeed because the content sounds like a real person. Machine learning-based algorithms can work around the familiar tone to spot unusual semantic patterns and formulaic scams.
- Machine learning can be used to detect malicious Twitter accounts that distribute spam, scams and misinformation. Researchers from the University of Iowa have developed a machine learning approach that can identify more abusive Twitter accounts faster and with a higher rate of accuracy.
- Finally, machine learning can be used to create cyber security baselines and identify situations that don’t match the anticipated patterns. People who work in cyber security are still the most critical part of securing organizations, but machine learning can help to herald when something might be wrong.
Machine learning to prevent cyber security breaches is exciting, but humans (you!) remain the best frontline defense. To learn more about how you can spot phishing emails, malicious scams and cyber security warning signs, visit the Security Matters blog and check out our resources for staff, students and faculty.
Source: Security Matters UoT