A Framework for BYOD Continuous Authentication: Case Study with Soft-Keyboard Metrics for Healthcare Environment

Mobile authentication is a hot topic because organizations can adopt BYOD (bring your own device) policies that allow to use  personal devices, ra-ther than require the use of officially provided devices. However, this brings additional access control issues like intentional or unintentional unauthorized uses of devices (e.g., stealing a mobile phone) that may eventually result in ac-cess to  sensitive information. Continuous authentication (CA) aims to mitigate and provide a solution to access control by monitoring user activity. CA can then be particularly useful in mobile BYOD environments. However, each CA solution has to be implemented and integrated ad-hoc and tailored for each par-ticular information system that wants to use it. This paper presents a modular, extensible framework for CA that enables to integrate new agents and models to implement access control with mobile devices. The framework includes three main types of components: Endpoint Detection and Response (EDR) Agents that run on the mobile device to gather user metrics and evaluate user’s trust, APIs that collect information and return trustworthiness levels of users, and AI models that predict the trust of users. The framework also integrates authorized third parties that can ask for trust levels of individual users and are responsible for implementing the resulting security measures like raising alerts. The archi-tecture is demonstrated in a healthcare environment which is part of the ProTe-go project. The proof-of-concept implements a mobile EDR agent and AI mod-el based on the soft-keyboard input data collected on the mobile phone.