Securing the web with AI

In today’s web not a week goes by without a large data breach or a website being hacked. Unlawful access to our online information can be prevented by encryption, 2FA and IP detection, but often this isn’t enough. Additionally tools like 2FA can be difficult to configure or understand for the initiated, leaving many at a disadvantage - security should be a right, not a privilege.

This session explores some interesting and experimental ways that AI could be used to improve security and protect users on the web.

Using existing AI tools like Azure Cognitive Services we’ll look at ways facial or voice recognition could be used as part of 2FA to make security more accessible to end users, including code examples in .NET and JavaScript. Then we’ll investigate some simple methods to find patterns in user behaviour, such as login times and mouse behaviour, to detect anomalies and take action.

Finally, we’ll see how machine learning can be used to identify fraudulent text or photoshopped images within webpages and warn users of scams before it becomes an issue.