Usable Secure Software Design and Development

点击翻译按钮获取中文摘要

Cybersecurity is now a problem of urgent global importance. While secure technical infrastructure remains critical, many successful and devastating attacks involve exploiting inevitable human fallibilities. Some attackers focus on deception affecting human decision-making, where end-users are tricked into trusting fraudulent email or websites, or installing malware. These lead to compromises in privacy and confidentiality, as well as theft of resources and ransomware. Other attackers focus on vulnerabilities in software code, where decades of work in software design and process do not address the adversarial nature of security, and programmers can easily overlook necessary checks or dependencies. My research program is in the general area of Usable Security, involving both Cybersecurity and Human-Computer Interaction. The program involves two specific lines of research. The first addresses end-user security decision-making. The second addresses the usable security issues that affect software developers. The work on security decision-making involves mental models. Mental model theory dates back decades, and supposes cognitive models of cause and effect that predict possible consequences. There are several related problem areas where this work is needed: website legitimacy, where attackers create fraudulent websites to deceive users to obtain credentials or provide malware; email legitimacy, where phishing email deceives users into visiting fraudulent websites or downloading malware; and software legitimacy, where attackers create malware positioned as benign software. All require users to make decisions, and in all cases users are poorly supported. I propose studying ways to design user interaction to better support and develop mental models of security, thus helping users make better security decisions. Another source of cybersecurity weakness involves software development, where accidental code vulnerabilities go undetected in software and supply chains. My particular interest has been on the collaborative aspects of software development, and I propose to extend this work to focus on collaborations relating specifically to security. Concepts for addressing security issues early within the development process have long been proposed, and seem to suggest collaboration can be very helpful. However, recent studies show these are little used and have low impact. My work will explore methods such adversarial collaboration and diverse fault detection, and I will work with software development teams to identify ways to improve.