Post on 21-May-2020
PRIVACY, BIG DATA AND GOVERNANCE MAY 2015
YOUR FACULTY
• Marty Abrams
• JoAnn Stonier
• Hilary Wandall
EVOLUTION OF AN INFORMATION ECONOMY IMPACTS COMPLIANCE AND GOVERNANCE
• Mainframe era – Data and systems were the same – early awareness
• Database technologies – Data could serve many systems, secondary
use became an issue – fair information practice principles based on
consent
• Emergence of predictive sciences – Broad data sets used to predict
future behavior – concerns about fairness and dignity
• Internet – Explosion of observation, often by many parties – new
ethics programs
• Internet of Things – Observation accelerates – growing tracking
discussions
• Big Data – All data available for predictions – governance becomes an
imperative
NATURE OF DATA RELATED TO INDIVIDUALS
• Origins of data have changed as technology has evolved
• Mainframe era data was collected directly from us
• Today fastest growing data sets come from observation and product of
predictive sciences
• To understand effective privacy one needs to understand the origins of
data
DATA TAXONOMY BASED ON ORIGIN
• Was developed for OECD March 2014 workshop
• Data may be segmented into four classifications based on
origin
• Provided – Comes directly from individual with their direct knowledge
• Observed – Individuals’ behavior is monitored on the Internet, public
places and sensors built into things like cars
• Derived – The product of grouping data and simple math
• Inferred – Categories that come from predicative sciences
POLICY IMPLICATIONS FROM TAXONOMY
• Observed and inferred data are fastest growing data sets
• Often without the attention of busy individuals
• Therefore consent, while very important, becomes less effective
as a governing mechanism
• Data originates in a manner invisible to individuals
• How does one fill the gap?
WHY IS THE BIG DATA DIFFERENT?
• Volume, velocity, diversity – blending data robustly
• More room for processing errors
• Diverse data sets stretch compatibility to stated purpose
• Goes beyond testing best intuition – data drives questions
• Business logic is a question
• Correlation not causation
• Is the correlation meaningful
SO WHY BIG DATA?
• Big data sees things that go beyond intuition, and that vision
improves outcomes
• Health
• Disaster relief
• Improved markets
• The improvements are compelling and inevitable
• Big data also may enhance discrimination and mistakes
• Governance must be in place to facilitate the compelling and
prevent inappropriate results
GOVERNANCE OBJECTIVES
• Determine when it is appropriate to repurpose data pertaining
to people
• Assure the processing is conducted in an accurate manner
• Guide the application of remarkable insights
IAF HAS BEEN DEVELOPING A GOVERNANCE PROCESS
• Based on key concepts
• Data protection assures the full range of fundamental rights, not just a
narrow definition of privacy
• Reticence risk is meaningful and real
• A fairness assessment is necessary
• This takes us beyond compliance to ethics
Part A
Part B
Part C
Part D
Big Data Project Structure
• Description of the unified ethical framework. • Creates a basis for the interrogation guidance.
• Interrogation guidance for implementing the code. • Illuminates the key issues that must be considered in
making a judgment on whether a Big Data project is fair, responsible and ethical
• Mechanisms for enforceability. • Ensures compliance with code.
• Contextual interrogation questionnaire. • Customized for organizations, industries, mediums.
FIND MARTIN ABRAMS
• mabrams@informationaccountability.org
• www.informationaccountability.org
• Key papers are on the website
• A number of those papers will soon be available in Japanese