Surveillance of Emergent Associations: Freedom of Association in a Network Society

advertisement
Surveillance of Emergent
Associations: Freedom of
Association in a Network
Society
Katherine J. Strandburg
DePaul University College of Law
OUTLINE
I.
A.
B.
C.
II.
A.
“NETWORK EFFECTS”:
EMERGENT ASSOCIATION
TRAFFIC DATA TRACKING
SOCIAL NETWORK ANALYSIS
THE PARADOX OF CURRENT LAW
SURVEILLANCE LAW’S NEGLECT OF
TRAFFIC DATA
B. FIRST AMENDMENT’S STRONG
PROTECTION OF ASSOCIATIONS
III. FOURTH AMENDMENT LESSONS FOR
FREEDOM OF ASSOCIATION
IV.
APPLICATIONS TO NETWORK ANALYSIS OF
TRAFFIC DATA
I.A. “NETWORK EFFECTS”:
EMERGENT ASSOCIATIONS





Internet +
Wireless technology +
Locational technology +
Social networks +
Video and audio technology
= New potential for association
- political, civic, social, educational, etc.
I. “NETWORK EFFECTS”
EMERGENT ASSOCIATIONS
 Can form rapidly and cheaply
 Link w/o rt distance, all sizes possible
 Strategies can self-organize
 Membership not specified or defined,
pseudonymous
ENORMOUS NEW POTENTIAL FOR
EXPRESSIVE ASSOCIATION
I.B.
“NETWORK EFFECTS”:
TRAFFIC DATA
BUT . . . .
 Takes place through communication
intermediaries
 Spontaneous association is recorded in
voluminous amounts of traffic data
- Telephone records
- Internet traffic logs
- Location tracking, etc.
 New technologies of association also used
by malevolent groups – terrorist, pedophiles
INCREASING RELATIONAL SURVEILLANCE
I.C. “NETWORK EFFECTS”:
NETWORK ANALYSIS
RELATIONAL SURVEILLANCE THROUGH
NETWORK ANALYSIS OF TRAFFIC DATA
Social Network Analysis and Network Science:
Use relational patterns to determine:
-Structure of organization
- Divide populations into social groups
- Understand roles played by different
individuals
- Using sophisticated data mining-type
algorithms and large datasets
I.C. NETWORK ANALYSIS
RELATIONAL SURVEILLANCE THROUGH
NETWORK ANALYSIS OF TRAFFIC DATA
 Investigating structure of known
networks
 Targeted link analysis
- investigate a “suspicious”
individual’s associational affiliations
Pattern-based analysis
- uncover associations using
relational patterns
- “match” to suspicious patterns
I. “NETWORK EFFECTS”
PROBLEMS W/ RELATIONAL
SURVEILLANCE USING NETWORK
ANALYSIS
 EXPOSE LEGITIMATE ASSOCIATIONS
 TARGETED LINK ANALYSIS EXPOSES
ASSOCIATIONS OF NON-TARGETS
 MANY PROBLEMS WITH ACCURACY
- Problems are deep, esp. for pattern
analysis, do “malevolent” associations really
look different?
- See many discussions of issues with data
mining (Slobogin, Swire)
OUTLINE
I.
A.
B.
“NETWORK EFFECTS”:
EMERGENT ASSOCIATION
RELATIONAL SURVEILLANCE
II. THE PARADOX OF CURRENT LAW
A.
B.
SURVEILLANCE LAW’S NEGLECT OF
TRAFFIC DATA
FIRST AMENDMENT’S STRONG
PROTECTION OF ASSOCIATIONS
III. FOURTH AMENDMENT LESSONS FOR
FREEDOM OF ASSOCIATION
IV.
APPLICATIONS TO NETWORK
ANALYSIS OF TRAFFIC DATA
II.A. PARADOX OF
CURRENT LAW
FOURTH AMENDMENT: virtually no protection to
traffic data
- Focus on individual privacy
-Third party doctrine
- Content/non-content distinction
- interception/stored records distinction
SURVEILLANCE STATUTES: only a little better
-Traffic data usually available by showing
“relevance” to a law enforcement or terrorism
investigation
- sometimes little or no judicial oversight
NETWORK ANALYSIS MAKES LOTS OF DATA
POTENTIALLY “RELEVANT”
II.B. PARADOX OF CURRENT LAW
FIRST AMENDMENT: strong protection for
expressive association
- Boy Scouts v. Dale: broad definition of
expressive association; deferential approach
to association perceptions of harm
- NAACP v. Alabama: compelled disclosure
of membership list requires strict scrutiny
under First Amendment
- Sheldon v. Tucker: broad, indiscriminate
disclosure of memberships cannot be
required; disclosure must be tailored to to
compelling gov’t interest
II. PARADOX OF CURRENT LAW
 Relational Surveillance Using Network
Analysis =
Disclosure Of Association Membership Lists,
Esp. For EMERGENT ASSOCIATIONS
 Relational Surveillance Threatens to Chill
Emergent (and other) Association b/c Fourth
Amendment fails to protect associational
information
 First Amendment has been applied only to
traditional associations
II. PARADOX OF CURRENT LAW
NEED FIRST AMENDMENT APPROACH TO
RELATIONAL SURVEILLANCE (CF. SOLOVE)
NEED TO UPDATE FREEDOM OF ASSOCIATION
DOCTRINE TO REFLECT CURRENT
ASSOCIATIONAL PATTERNS AND
DISCLOSURE MECHANISMS
OUTLINE
I.
A.
B.
“NETWORK EFFECTS”:
EMERGENT ASSOCIATION
RELATIONAL SURVEILLANCE
II. THE PARADOX OF CURRENT LAW
A.
B.
SURVEILLANCE LAW’S NEGLECT OF
TRAFFIC DATA
FIRST AMENDMENT’S STRONG
PROTECTION OF ASSOCIATIONS
III. FOURTH AMENDMENT LESSONS FOR
FREEDOM OF ASSOCIATION
IV.
APPLICATIONS TO NETWORK
ANALYSIS OF TRAFFIC DATA
III. FOURTH AMENDMENT
LESSONS
HOW TO ADAPT TO EVOLVING
TECHNOLOGY ?
 Law must adapt to change of situs of
socially important activity (Katz)
 Law must recognize that surveillance
involves both gathering and analyzing
data – both can implicate Constitutional
values (Kyllo)
 Intrusiveness of surveillance depends on
discrimination between legitimate and
illegitimate activity (“dog sniff” cf. Kyllo)
III. FOURTH AMENDMENT
LESSONS
FREEDOM OF ASSOCIATION TEST for
RELATIONAL SURVEILLANCE
- Does the surveillance serve a compelling
interest?
- Is analysis sufficiently accurate in light of
potential burden on association?
- Is analysis sufficiently closely related to
the compelling interest in light of potential
burden on association?
OUTLINE
I.
A.
B.
“NETWORK EFFECTS”:
EMERGENT ASSOCIATION
RELATIONAL SURVEILLANCE
II. THE PARADOX OF CURRENT LAW
A.
B.
SURVEILLANCE LAW’S NEGLECT OF
TRAFFIC DATA
FIRST AMENDMENT’S STRONG
PROTECTION OF ASSOCIATIONS
III. FOURTH AMENDMENT LESSONS FOR
FREEDOM OF ASSOCIATION
IV.
APPLICATIONS TO NETWORK
ANALYSIS OF TRAFFIC DATA
IV. APPLICATIONS TO NETWORK
ANALYSIS
Targeted Link Analysis:
 Involves traffic data from target plus
connected individuals
 For Target: exposes numerous associations
indiscriminately
- Cf. Sheldon v. Tucker – high standard
should be required (probable cause?)
 For others: less intrusive, but also less
relevant (depending on how many links away)
- Relevance not enough
- Reasonable suspicion?
IV. APPLICATIONS TO NETWORK
ANALYSIS
Pattern-Based Analysis:
 Involves traffic data from numerous individuals
 Exposes broad swath of associations (mostly
innocent)
 Ability to distinguish malevolent associations
questionable at best
DOES NOT MEET FIRST AMENDMENT
STANDARDS –
NOT SUBSTANTIALLY RELATED
IV. CONCLUSIONS
 Associations poised to play increasingly
important role in democracy and culture b/c of
Internet, etc.
 Traffic data increasingly permits relational
tracking
 Fourth Amendment fails to protect traffic data
b/c focus is on individual privacy
 Freedom of association strongly protects
traditional associations
 Need to update freedom of association law to
regulate relational surveillance using traffic data
IV. CONCLUSIONS
 Pattern-based network analysis:
-Does not meet 1st Amendment standards
Targeted Link Analysis
-Should require probable cause for target
- Reasonable suspicion for “links”
 Individual traffic data
-Case by case analysis
- some cases may implicate First
Amendment prohibition
Download