Type of Talk: UTCS Colloquium

Speaker/ Affiliation: Ar

vind Narayanan/ Stanford University

Date/Time: Thursday, November 5

, 2009/ 11:00 a.m.

Location: ACES 2.402

Host: Vitaly Shmatikov<

/p>

Talk Title: "What does it mean to own our genes?"

T

alk Abstract:
Given that each of us shares genetic material with our b

lood
relatives, to what extent can one expect to keep one''s genetic<

br />information private? I will consider this question with respect to an<

br />attacker equipped with large-scale (albeit incomplete and "noisy

")
information about the blood relationships in a large populati

on group,
i.e., a genealogical graph.

Given this kind of

auxiliary information, it turns out that the
availability of genotype
information of a small fraction of
individuals -- as little as 0.2%,
in preliminary experiments -- is
enough to cause the majority of the

population to lose any hope of
genetic privacy. I will describe a stro

ng inference attack that allows
the attacker to re-identify completely
anonymous genetic material,
such as pieces of hair collected en mass

e from public spaces without
the consent or even the knowledge of the

potential victims.

There are many ongoing efforts aimed at aggre

gating genealogical data
on a massive scale. As I will show, the comp

ilation of the "world''s
family tree" is a matter of time.
Further, there are several
population groups for which enough auxili

ary data is already available
to leave them vulnerable to genetic re-i

dentification.

There is no purely technological fix to this atta

ck. I will briefly
present policy prescriptions that may delay ubiquit

ous genetic
re-identifiability, and argue that genetic privacy norms

must change
to accommodate the new technological reality.

S

peaker Bio:
Arvind Narayanan is a post-doctoral researcher at Stanford

. He
recently finished his Ph.D at the University of Texas at Austin,

advised by Vitaly Shmatikov. His research is on the privacy and

anonymity issues involved in publishing large-scale datasets about
peo

ple. His thesis, in a sentence, is that the level of anonymity that

society expects - and companies claim to provide - in published databases i

s fundamentally unrealizable. He blogs about his anonymity-breaking efforts
and other research at http://33bits.org/