© Fraunhofer ISI
Daniel Bachlechner, Fraunhofer ISI
13th Pre-ICIS Workshop on Information Security and Privacy 13 December 2018
PRIVACY-PRESERVING TECHNOLOGIES
IN THE AGE OF BIG DATA
© Fraunhofer ISI
Page 2
e-SIDES helps making big data solutions and
applications more responsible
Identify
ethical and
societal issues
Identify PPTs
Assess
existing
technologies
Conduct a
gap
analysis
Make
recom-
menda-
tions
Identify
design
require-
ments
Identify
imple-
mentation
barriers
© Fraunhofer ISI
Page 3
Agenda
 Status quo
 Research question
 Methods
 Results
 Summary
 Contribution
 Outlook
© Fraunhofer ISI
Page 4
Big data applications have positive and
negative effects
Big data
applications
Negative effects
 Intrusion into peoples' privacy
 Unfair treatment of people and
organizations
 Dependence on organizations and
technology
 Lack of decision transparency
Positive effects
 Faster, fact-based decision-making
 Increasing possibilities of
automation
 Emergence of new and improved
products and services
 Cost savings
© Fraunhofer ISI
Page 5
PPTs can be assigned to different classes of
technologies
PPTs
Approaches
Anonymization and
sanitization
 k-anonymity
 Differential privacy
 Randomization
Encryption and multi-party
computation
Access control and policy
enforcement
Accountability, transparency
and data provenance
Access, portability and user
control
 Attribute-based encryption
 Homomorphic encryption
 Attribute-based access control
 Sticky policies
 Automated auditing processes
 Multi-channel and layered
approaches
 Standard exchange formats
 Privacy preferences
 Personal data stores
© Fraunhofer ISI
Page 6
Big data applications
The role of PPTs in the age of big data is not
fully understood
Big data solutions
?PPTs
 Effectiveness
 Challenges
 Perception
 Use
© Fraunhofer ISI
Page 7
PPTs
PPTs were assessed though interviews and a
review of related literature
Semi-structured
interviews
 Nine key informant
interviews
 Researchers, company
representatives,
members of relevant
organizations
 30-45 minutes
 16 short interviews
 Mainly researchers
 <10 minutes
Literature review
Technology-specific
assessment
General assessment
 Technology
deployment
 Demand for privacy
protection
 Regional differences
 Technological
boundaries
 Societal responsibility
© Fraunhofer ISI
Page 8
PPTs differ in term of their effectiveness and
the challenges they pose
Challenges
 Failure of anonymization
 Unique characteristics
 Mathematical background stable
Anonymization
and sanitization
Encryption and
MPC
Access control and
policy
enforcement
Accountability,
transparency and
data provenance
Access, portability
and user control
Effectiveness
 Privacy-utility trade-off
 Choice of parameters
 Adversary’s knowledge unknown
 Untrusted parties
 Computation cost
 Ease of use
 Single point of failure
 Opacity
 Increasing complexity
 Explanation of decisions
 Heterogeneity of data
 Large amounts of information
 Different protection levels
 Need for prior agreements
 Privacy by default
 Encryption is strong
 Relatively slow
 Limited availability of tools
 Low granularity and flexibility
 Automated enforcement
 Boundaries between jurisdictions
 Machine learning modules
 Most devices do not have screens
 Implementation of fixes
 Untrusted parties
 Informational asymmetry
 Shift of responsibility
© Fraunhofer ISI
Page 9
PPTs show similarities in terms of their
perception and use
PPTs
Perception and use
Technology
deployment
 Minimal integration of PPTs into solutions
 Fear of economic disadvantages
 Privacy by design has not arrived in practice
Demand for
privacy protection
Regional
differences
Technological
boundaries
Societal
responsibility
 Little demand for solutions that include PPTs
 Privacy protection as a selling point
 Lack of risk awareness
 Considerable regional differences
 Local legislation affects technology globally
 Broad access to technologies may be restricted
 Limitations of reactive approaches
 Need for awareness and education
 Make sure technology functions as expected
 Strengths of the parties is relevant
 Raise awareness and exercise rights
 PPT integration alone would not be sufficient
© Fraunhofer ISI
Page 10
PPTs are essential but their use is only slowly
becoming common practice
 Status quo
 Research question
 Methods
 Results
 Summary
 Contribution
 Outlook
 Big data applications have positive and
negative effects
 PPTs can be assigned to different classes
of technologies
The role of PPTs in the age of big data is
not fully understood
PPTs were assessed though interviews and a
review of related literature
 PPTs differ in term of their effectiveness
and the challenges they pose
 PPTs show similarities in terms of their
perception and use
© Fraunhofer ISI
Page 11
Big data applications benefit from PPTs but
individual circumstances must be considered
Big data applications
Big data solutions
?PPTs
 PPTs pursue different
aims and thus need to
be combined
 No single most
important class of
technologies
 A multidimensional
measure for privacy
protection is needed
 Objectives of big data
and privacy seem
almost antagonistic
© Fraunhofer ISI
Page 12
Recommendations for developers and users of
big data solutions will be made
Big data applications
Big data solutions
PPTs
 Conduct a gap analysis
 Identify design
requirements
 Identify implementation
barriers
 Make recommendations
© Fraunhofer ISI
Page 13
If you would like to contribute to the
research, please get in touch
https://e-sides.eu/
info@e-sides.eu
© Fraunhofer ISI
Page 14
If you would like to get in touch with me
Dr. Daniel Bachlechner
Competence Center Emerging Technologies
Fraunhofer Institute for Systems and Innovation Research
Breslauer Strasse 48 | 76139 Karlsruhe | Germany
Phone +49 721 6809-161 | Fax +49 721 6809-315
daniel.bachlechner@isi.fraunhofer.de

e-SIDES presentation at WISP 2018, San Francisco 13/12/2018

  • 1.
    © Fraunhofer ISI DanielBachlechner, Fraunhofer ISI 13th Pre-ICIS Workshop on Information Security and Privacy 13 December 2018 PRIVACY-PRESERVING TECHNOLOGIES IN THE AGE OF BIG DATA
  • 2.
    © Fraunhofer ISI Page2 e-SIDES helps making big data solutions and applications more responsible Identify ethical and societal issues Identify PPTs Assess existing technologies Conduct a gap analysis Make recom- menda- tions Identify design require- ments Identify imple- mentation barriers
  • 3.
    © Fraunhofer ISI Page3 Agenda  Status quo  Research question  Methods  Results  Summary  Contribution  Outlook
  • 4.
    © Fraunhofer ISI Page4 Big data applications have positive and negative effects Big data applications Negative effects  Intrusion into peoples' privacy  Unfair treatment of people and organizations  Dependence on organizations and technology  Lack of decision transparency Positive effects  Faster, fact-based decision-making  Increasing possibilities of automation  Emergence of new and improved products and services  Cost savings
  • 5.
    © Fraunhofer ISI Page5 PPTs can be assigned to different classes of technologies PPTs Approaches Anonymization and sanitization  k-anonymity  Differential privacy  Randomization Encryption and multi-party computation Access control and policy enforcement Accountability, transparency and data provenance Access, portability and user control  Attribute-based encryption  Homomorphic encryption  Attribute-based access control  Sticky policies  Automated auditing processes  Multi-channel and layered approaches  Standard exchange formats  Privacy preferences  Personal data stores
  • 6.
    © Fraunhofer ISI Page6 Big data applications The role of PPTs in the age of big data is not fully understood Big data solutions ?PPTs  Effectiveness  Challenges  Perception  Use
  • 7.
    © Fraunhofer ISI Page7 PPTs PPTs were assessed though interviews and a review of related literature Semi-structured interviews  Nine key informant interviews  Researchers, company representatives, members of relevant organizations  30-45 minutes  16 short interviews  Mainly researchers  <10 minutes Literature review Technology-specific assessment General assessment  Technology deployment  Demand for privacy protection  Regional differences  Technological boundaries  Societal responsibility
  • 8.
    © Fraunhofer ISI Page8 PPTs differ in term of their effectiveness and the challenges they pose Challenges  Failure of anonymization  Unique characteristics  Mathematical background stable Anonymization and sanitization Encryption and MPC Access control and policy enforcement Accountability, transparency and data provenance Access, portability and user control Effectiveness  Privacy-utility trade-off  Choice of parameters  Adversary’s knowledge unknown  Untrusted parties  Computation cost  Ease of use  Single point of failure  Opacity  Increasing complexity  Explanation of decisions  Heterogeneity of data  Large amounts of information  Different protection levels  Need for prior agreements  Privacy by default  Encryption is strong  Relatively slow  Limited availability of tools  Low granularity and flexibility  Automated enforcement  Boundaries between jurisdictions  Machine learning modules  Most devices do not have screens  Implementation of fixes  Untrusted parties  Informational asymmetry  Shift of responsibility
  • 9.
    © Fraunhofer ISI Page9 PPTs show similarities in terms of their perception and use PPTs Perception and use Technology deployment  Minimal integration of PPTs into solutions  Fear of economic disadvantages  Privacy by design has not arrived in practice Demand for privacy protection Regional differences Technological boundaries Societal responsibility  Little demand for solutions that include PPTs  Privacy protection as a selling point  Lack of risk awareness  Considerable regional differences  Local legislation affects technology globally  Broad access to technologies may be restricted  Limitations of reactive approaches  Need for awareness and education  Make sure technology functions as expected  Strengths of the parties is relevant  Raise awareness and exercise rights  PPT integration alone would not be sufficient
  • 10.
    © Fraunhofer ISI Page10 PPTs are essential but their use is only slowly becoming common practice  Status quo  Research question  Methods  Results  Summary  Contribution  Outlook  Big data applications have positive and negative effects  PPTs can be assigned to different classes of technologies The role of PPTs in the age of big data is not fully understood PPTs were assessed though interviews and a review of related literature  PPTs differ in term of their effectiveness and the challenges they pose  PPTs show similarities in terms of their perception and use
  • 11.
    © Fraunhofer ISI Page11 Big data applications benefit from PPTs but individual circumstances must be considered Big data applications Big data solutions ?PPTs  PPTs pursue different aims and thus need to be combined  No single most important class of technologies  A multidimensional measure for privacy protection is needed  Objectives of big data and privacy seem almost antagonistic
  • 12.
    © Fraunhofer ISI Page12 Recommendations for developers and users of big data solutions will be made Big data applications Big data solutions PPTs  Conduct a gap analysis  Identify design requirements  Identify implementation barriers  Make recommendations
  • 13.
    © Fraunhofer ISI Page13 If you would like to contribute to the research, please get in touch https://e-sides.eu/ info@e-sides.eu
  • 14.
    © Fraunhofer ISI Page14 If you would like to get in touch with me Dr. Daniel Bachlechner Competence Center Emerging Technologies Fraunhofer Institute for Systems and Innovation Research Breslauer Strasse 48 | 76139 Karlsruhe | Germany Phone +49 721 6809-161 | Fax +49 721 6809-315 daniel.bachlechner@isi.fraunhofer.de