## Security for Humans Thinking back over the semester --- ## Making Security Usable Security must fit with four key elements: 1. Capabilities and limitations of the target users 2. Goals of the users * Security is a enabling task 3. Physical and social context of use 4. Capabilities and limitations of the device used ### Users are not the enemy --- ## How to Design for Security * Largely, same as other UX situations * Contextual inquiry: Talk to users! * Users don't tell you if it is secure * Enabling task: Can't just watch them easily --- ## Ethics of Studying Users for Security * Informed Consent * Deception? * Attacking Users? --- ## Authentication: Which User Is It? * Ask the user to do something that only they can do * Usablility: They do it a lot * Security: No one else can do it * Factors: Something you know, have, or are --- ## Authentication: Which User Is It? * Passwords are the most common * Passwords have a "life cycle" * Users adjust passwords based on their security/usability needs * Biometrics are usable? secure? * Multiple factors helps --- ## Access Control: What is the User Allowed to Do? * Permissions grant *future* access, but changing is hard * Lots of permissions are overwhelming * Limit options * When to ask? --- ## Secure Communication: Understanding Encryption * Most users don't understand encryption * Interfaces need to communicate *enough*, without overwhelming * Technical cryptography constraints pose challenges --- ## Warnings: Designing Communication * Risky situation, potential consequences, instructions * To dangerous to not warn, too safe to prohibit * Third line of defense (design away, guard against) --- ## Warnings: Designing Communication * NEAT Warnings: Necessary, Explained, Actionable, Tested * SPRUCE explanations * Design of communication really shapes user actions (e.g. SSL warnings) --- ## Warnings: Designing Communication * Choosing indicators is hard, but important * Good indicators convey a LOT in little space * Chrome URL bar * Lots of indicators: a nutrition label approach * Support good decisions --- ## Nudges: Using Design to Influence Decisions * Controlling **Information** * **Presentation** and framing * Setting **Defaults** * Changing **Incentives** * **Reversibility** and error resiliency * Controlling **Timing** --- ## Stopping Phishing: A User Security Issue * Very difficult to use design to stop * If we know something is phishing, just block it! * Simulated phishing campaigns? * Training and education? --- ## At-Risk Users * Some people / situations produce more security risk * Changes decision-making, design constraints * Sometimes, **very** serious * Mostly, follow UX design methods * **Don't blame the user** --- ## Core Ideas * Every system needs security designed in * Security is a secondard task! * Users care about security, even if they can't do it * Don't blame the user * Design from data, and security * Don't (try to) make users into experts * Design + Communication = Security