Turning Privacy Risk Assessment Into 20 Questions for Developers
summary
Prototype PrivacyAkinator converts privacy design into LLM-generated multiple-choice flows that prompt roughly 20 closed-ended questions, using a verb-based representation and a taxonomy mined from 10,000 privacy news articles. It asks concrete options like opt-in / opt-out / always-on, updates a data-action graph dynamically, and was evaluated with 24 students/junior developers where it finds key decisions at 98% vs 42% coverage compared to a NIST PRIME baseline and reports lower mental workload.
AI-written summary · about
chapters (8)
- 0:00Privacy risk assessment intro
- 1:57Privacy Actuator question gen
- 4:45Challenges with frameworks
- 6:54Designing Privacy Actuator
- 9:01Verb-based privacy rep
- 11:50Dynamic privacy questions
- 12:53User evaluation of tool
- 13:58Next steps and research
related talks
11mEvaluating the Usability of Privacy Choice Mechanisms
20mPrivacy Shift Left: A Machine-Assisted Threat Modeling Approach
51mFilling the gap between DPOs and developers
10mEvaluating Privacy Perceptions, Experience, and Behavior of Software Development Teams
16mAutomating Technical Privacy Reviews Using LLMs
05mDeciding on Personalized Ads: Nudging Developers About User Privacy
40mCreative Ways to Champion Privacy Engineering Awareness
01h 16mCommercial opportunities for designing privacy in: Embedding privacy ...