Research Overview

1
Research Overview
2
Research Overview

• VR
• Online Communities
• Communication
• Collaboration
• Olympic Games
• Privacy

3
Virtual Reality

• Virtual reality
• Designing physical environments
• Simulator sickness

4
Social Dilemma Experiments

• Tension between common good and individual greed
• Group benefits by collaborating
• Individual benefits by tricking others
• Examine how different environments affect
  collaboration and trust

5
Social Dilemmas Effect of Communication Modality
The Effect of Communication Modality on
Cooperation in Online Environments Proceedings
of ACM Conference on Human Factors in Computing
Systems CHI 2000
6
Social Reputation Systems - Profiles
Social Networks
Social network
User profile
Relation to me
Finding Others Online Reputation Systems for
Social Online Spaces Proceedings of ACM
Conference on Human Factors in Computing Systems
CHI 2002
7
Olympic Games

• Olympic Games
• Supporting large organizations
• HCI in real-time/ critical systems
• Commentator Information System
• Getting free tickets

8
Research Overview

• VR
• Online Communities
• Communication
• Collaboration
• Olympic Games
• Privacy

9
Motivation

• Privacy Security in part a awareness and
  management problem
• Users need help to make informed and correct
  decisions
• Users largely unaware of privacy and security
  issues something bad happens
• Even when they do get involved, interfaces are
  typically inadequate

10
Challenges in Privacy-Aware Design

• Two lines of research
• Understand users expectations, mental models,
  and the tradeoffs users they make with regards to
  personal privacy
• Understand design developing methods to
  structure the identification and analysis of
  privacy vulnerabilities as part of the design
  process

11
Privacy Surveys Experiments

• Observe users as well as ask for self-ratings
• Force tradeoffs, find priorities
• Components
• Demographic
• Privacy questions
• Knowledge challenge
• E-commerce experiment

• 175 volunteer subjects completed whole experiment
• Population
• Predominantly male (74)
• Predominantly US (95)
• More educated than average
• High computer expertise
• Normal mix of concerns

Carlos Jensen, Colin Potts, Christian Jensen.
Privacy Practices of Internet Users Self-report
versus Observed Behavior International Journal
of Human Computer Studies special issue on HCI
Research in Privacy and Security. October 2005.
12
Knowledge Challenge

• What do people actually know about online
  privacy?
• People are very overconfident
• Especially for well-known technologies
• Survey results must be taken with a grain of salt
• Actual knowledge of risks abysmal

13
E-commerce Experiment

• We know what people say they look at before
  buying online
• Does observed behavior match stated behavior?
• What are the most influential factors?
• How strong are these priorities?
• Force tradeoffs, find priorities

14
(No Transcript)
15
E-commerce Experiment

• We know what people say they look at before
  buying online
• Does observed behavior match stated behavior?
• What are the most influential factors?
• How strong are these priorities?
• Force tradeoffs, find priorities

• 12 Variables
• 20 Discount
• Credit Card Symbols
• SSL
• Blocked Cookies/P3P
• Contact Telephone
• Contact Address
• Contact Email
• Privacy Seal
• 4 privacy policies
• User oriented Good
• User oriented Bad
• Company oriented Good
• Company oriented Bad
• 66 possible combinations
• 1200 observations

16
Experimental Results

• Sub-populations differ in their models
• Gender
• Westin privacy segmentation
• Experience
• Why is this useful?
• Tells us a what people look for
• Identify perception gap
• Lessons for UI design
• Users looking for trust markers
• Trust policies, but dont read them
• Prone to wrong assumptions
• Indifferent to important markers

17
iWatch Privacy Management

• Web privacy
• Scarce information
• Lack of awareness
• Lack of options
• Putting users in charge
• Own evaluation
• Own filtering/blocking
• Two components
• Web Crawler
• Proxy/UI

Central Server
User PC
18
iWatch Crawler

• Index 1 million web-pages/week
• 180,000 unique domains
• Seeded with most visited sites
• Indexes on 21 criteria
• Web-bugs / 3rd party resources (187,000)
• Cookies (395,000)
• P3P policies and mismatches (41,500)
• Track history of changes
• Detect trends impact of regulation/court
  action/other
• Detects information sharing networks

19
(No Transcript)
20
Top 20 3rd Party-Resource Servers

• graphics.fansonly.com
• graphics.ocsn.com
• server-au.imrworldwide.com
• i.cmpnet.com
• welcome.hp-ww.com
• www.qksrv.net
• dw.com.com
• pics.ebaystatic.com
• server-dk.imrworldwide.com
• ehg-dig.hitbox.com

• server-us.imrwordwide.com
• pagead2.googlesyndication.com
• switch.atdmt.com
• hostingprod.com
• ehg-findlaw.hitbox.com
• view.atdmt.com
• incude.ebaystatic.com
• images.findlaw.com
• ads.api.no
• ad.doubleclick.net

Of the top 20 3rd party-resource servers 70 are
ad-servers or marketers 30 are dedicated image
or script servers
21
pics.ebaystatic.com include.ebaystatic.com
Web-bugs Ad-servers
graphics.fansonly.com graphics.ocsn.com
22
iWatch Client

• Crawler data interesting, but overwhelming
• Not interested to users
• Not usable
• Survey says users want simple ratings
• Derive risk profiles/scores
• Make things visible
• Informing without distracting
• User care about privacy, but not primary goal

23
iWatch Client Prototype

• How much information is appropriate?
• How to present that information
• What settings do people settle on?
• Basic information for how to build future UIs

24
Current/Future

• Olympic Commentator Information System
• Privacy Awareness
• Open Source Software Usability