Transcript too much information?: information privacy and

TOO MUCH INFORMATION?
PRIVACY, RATIONALITY, AND
TEMPTATION
Katherine J. Strandburg
DePaul University College of Law
MOTIVATION FOR
INFORMATION PRIVACY
AUTONOMY AND DIGNITY FOR ME
v.
FREE FLOW OF INFORMATION FOR YOU
(BETTER DECISIONS)
?
IS MORE INFORMATION
ALWAYS BETTER?
SOCIAL NORMS AGAINST DISCLOSING
PERSONAL INFORMATION SUGGEST
THE ANSWER IS NO
WHY DON’T WE WANT TO LEARN
EVERYTHING ABOUT EVERYONE?
TOO MUCH INFORMATION?
TOO MUCH INFORMATION CAN DEGRADE
DECISIONMAKING IN 3 WAYS:
1)
MORE INFO NOT ALWAYS BETTER IF IT IS
MISLEADING (EVEN IF COMPLETELY
RATIONAL)
2)
BOUNDED RATIONALITY – “MISTAKES” IN
ANALYSIS DUE TO LOSS AVERSION,
INACCURATE PROBABILITY ASSESSMENT,
MENTAL ACCOUNTING
3)
LIMITED WILLPOWER – INABILITY TO IGNORE
IRRELEVANT INFORMATION (“MORE
PREJUDICIAL THAN PROBATIVE”)
TOO MUCH INFORMATION?
CONCLUSION:
CONTROLLING FLOW OF PERSONAL
INFORMATION IS A TWO-WAY STREET: BOTH
SUBJECTS AND RECIPIENTS MAY BENEFIT
FROM PRIVACY
BOTH SUBJECTS AND RECIPIENTS CAN BE
AFFECTED BY BOUNDED RATIONALITY
BOTH SUBJECTS AND RECIPIENTS CAN SUFFER
FROM SELF-CONTROL FAILURES
IMPLICATIONS for
PERSON-TO-PERSON
DISCLOSURES
• SOCIAL NORMS REGULATE INFORMATION
BY CONTEXT, “APPROPRIATENESS”
(ABILITY TO PROCESS ACCURATELY)
• PRIVACY TORTS SHOULD DO THE SAME
BUT WHAT ABOUT COMPUTERIZED DATA
PROCESSING?
CAN COMPUTERS HAVE
TOO MUCH INFORMATION?
• WHO DECIDES WHAT INFORMATION
TO COLLECT?
- Loss aversion of agents
- Private rationality v. Social optimality
- Time-inconsistent preferences of agents
(short term v. long term business goals)
- Externalities on subjects of information
CAN COMPUTERS HAVE
TOO MUCH INFORMATION?
• WHO DECIDES WHAT THE COMPUTER
DOES?
- Statistical optimization techniques (such as
data mining) rely on human input
- What data is “significant”
- When are results “significant”, “interesting”
- Statistical optimization techniques may be
“distracted” into local optima
- Categorical choices about “relevance” of
information may avoid some pitfalls
CAN COMPUTERS HAVE
TOO MUCH INFORMATION?
• WHO ASSESSES THE OUTPUT?
- Tendency to over-emphasize the certainty of
quantitative output
- Tendency to over-emphasize specific output
- Tendency to inaccurately interpret statistical
output
CAN COMPUTERS PROVIDE
TOO MUCH INFORMATION?
• COMPUTERIZED AGGREGATION OF DATA
PERMITS TARGETED MARKETING
- Targeted provision of information?
OR
- Targeted persuasion?
UNDERMINING STRATEGIC IGNORANCE?
ENTRAPMENT?
CAN COMPUTERS PROVIDE
TOO MUCH INFORMATION?
- COMPUTERIZED AGGREGATION OF DATA
MOVES INFORMATION BETWEEN SOCIAL
GROUPS
- INTERFERING WITH SOCIAL NORMS OF
APPROPRIATE DISCLOSURE?
- NORMS THAT WE PROTECT WITH THE
PRIVACY TORTS?
CONCLUSION
- THE SOCIAL VALUE OF “FREE FLOW OF
INFORMATION” PERSONAL INFORMATION
MAY BE LESS THAN IT SEEMS
- HUMAN COGNITIVE AND WILLPOWER
LIMITATIONS SHOULD BE KEPT IN MIND
- THE VALUE AND MEANING OF
COMPUTERIZED DATA PROCESSING ALSO
DEPENDS ON UNDERLYING HUMAN INPUT,
ANALYSIS, AND RESPONSE