PriBots Conversational Privacy with Chatbots Hamza Harkous 1 , Kassem Fawaz 2 , Kang. G. Shin 2 , Karl Aberer 1 1 EPFL; 2 University of Michigan Workshop on the Future of Privacy Notices and Indicators, SOUPS 2016
Privacy Notice: Current State 2
Privacy Notice: Current State legally binding Dual Role: human understandable 2
Privacy Notice: Current State legally binding Dual Role: human understandable Can we split these roles? 2
Standardization 3
Standardization Summarization 3
Standardization Summarization Challenges one size fits all user education 3
Privacy Choice: Current State 4
Privacy Choice: Current State fragmented ecosystem di fg icult to find 4
Conversation-first Interfaces 5
The Rise of Conversational UI 6
PriBots: Conversational Privacy Bots | Message 7
PriBots: Conversational Privacy Bots | Message 7
PriBots: Conversational Privacy Bots Appeal to new tech adopters | Message 7
PriBots: Conversational Privacy Bots Appeal to new tech adopters Appeal to existing users | Message 7
PriBots: Conversational Privacy Bots Appeal to new tech adopters Appeal to existing users | Message An intuitive way to 1. communicate privacy policies 2. adjust privacy preferences 7
1- Communicating Privacy Policies 8
Channel 9
Channel Primary 9
Channel Primary Secondary 9
Timing 10
Timing At-setup 10
Timing At-setup On-demand 10
k c a b d e e F implicit : sentiment analysis explicit : structured messages gathering users’ concerns 11
Voicing User Concerns Providers traditionally say what they want 12
Voicing User Concerns Providers Users’ concerns traditionally might not say what they be covered want 12
Voicing User Concerns Providers Users’ concerns PriBots traditionally might not activate the say what they be covered two-way channel want 12
2- Setting Privacy Preferences 13
14
Service and platform-dependent interface 14
Service and platform-dependent interface Tradeo fg s for simplicity: try finding this setting on Mobile Web version 14
15
Unique interface with all functionalities Ability to suggest adjustments to the user (combining notice and choice/preferences ) 15
System Architecture User Analysis & Input Classification 16
System Architecture Yes User Analysis & Structured Query Input Classification Query No Statement 16
System Architecture Yes User Analysis & Structured Retrieval Query Result Input Classification Query Module No Statement Knowledge Base 16
System Architecture Answer Yes Yes User Analysis & Structured Retrieval Query Confident? Result Formulation Input Classification Query Module in NL No Fallback Statement No Answer Knowledge Generation Base 16
System Architecture Answer Yes Yes User Analysis & Structured Retrieval Query Confident? Result Formulation Input Classification Query Module in NL No PriBot Fallback Statement No Reply Answer Knowledge Generation Base 16
System Architecture Answer Yes Yes User Analysis & Structured Retrieval Query Confident? Result Formulation Input Classification Query Module in NL No PriBot Fallback Statement No Reply Answer Feedback Knowledge Feedback Generation DB Base DB Augment the Knowledge Base Amendments/ Analytics Improvements • unanswered queries • frequent questions • user sentiments 16
Challenges 17
Mature User Understanding Text processing Question answering Domain-specific datasets and ontologies Graceful fallback 18
Legal Challenges Inherently error prone: are they legally binding? Accounting for false-positives and false-negatives The case of 3rd party PriBots: defamation possibilities? 19
Trusting the Machine rule-based vs. AI-based user backlash? regulate the confidence level 20
PriBots’ Personality positive tone → higher trust diversified content → reduced habituation 21
t n e m y o l p e D 22
t n e m y o l p e D provider 3rd parties 22
t n e m y o l p e D provider 3rd parties Suitable for Voice Assistants 22
What’s Next? Privacy as a Rule-based System User Dialogue Prototype Implementation studies 23
Questions/Feedback? hamza.harkous@gmail.com hamzaharkous.com 24
Image/Media Credits Zara Picken: slide 12 Egor Kosten: slide 24 Alex Prokhoda: slide 6 Freepik: slide 23 Geo fg Keough: slide 14 Victor: slide 12
Recommend
More recommend