Development of a networked behavioural-biometric authentication model

For anonymous, online transactions today, we typically use authentication in the form of knowledge “something you know”. This knowledge doesn’t need trust, because it’s binary: you know the secret passphrase (or possess second factor) or you do not. The authentication model is struggling, and, as a result, our identity is under attack: 1.5bn user records were stolen last year alone. Our authentication credentials have value to hackers and the data loss also has a substantial cost to business (a situation likely to worsen for various reasons). In contrast, recognition, like trust, is, instead, variable, and can be used to detect fraudsters. The project takes the existing log-in protocol we are all familiar with, and – without inconvenience – adds behavioural biometric (taps, swipes, keystrokes etc.), plus other more traditional, features to data capture. Those features draw a picture of the customer to generate a trust score, in real-time, for business to use in its online fraud detection efforts.