Content uploaded by Michele Loi
Author content
All content in this area was uploaded by Michele Loi on Jun 12, 2018
Content may be subject to copyright.
This is only a summary version of the paper, including the introduction and
conclusion (full text) and a bulletpoint summary of the other sections. Courtesy of
Springer, the full paper is provisionally available here https://rdcu.be/V1qH
The digital phenotype: a philosophical and ethical exploration
!
"
#
Introduction
$%&'()*+
$% &'()*
$, &'(-* !
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
#(
.
/
#
&0
# $
*"
1
!(!"
1 !234
.&'(-5$#4*62#
.#
78
/+
2bots$*
$&'(9*#:,+;1:
0"
;7.,:77
<!;Information, Communication & Society())
$%(&'(&*1=)'
2
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
.+ !
#
!
! & !
1
#
!
#>!?!
#
+ 1
digital
phenotype
#
1. What is the digital phenotype?
3
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
Figure 1. Analogies between the extended phenotypes of spiders and humans. Built with Apple
Keynote 8.0.1, except for web image, by Denis Frezzato, from The Noun Project, CC BY 3.0,
https://commons.wikimedia.org/w/index.php?curid=24126123
Organisms evolve together with the environments they shape or create. E.g.
beehives, beaver dams.
Digital data is analogous to beaver dams or beehives in humans: these are all
collectively created extended phenotypes involved in evolutionary feedback loops
(Richard Dawkins).
Human evolution is also cultural. Data affects the beliefs, norms and behavior in the
human population, thus affecting the cultural dimension of human evolution, most
directly (effects on other levels of evolution cannot be excluded in principle).
The metaphor of data as “digital footprint” is misleading. Rather data for humans is
like the beehive for bees, not a passive trace left somewhere, but something that
groups create collectively and in a way that influences the further behavior of its
producers
2. The ethics of the digital phenotype
2.1 The limits of the personal data protection approach
The data protection approach protects the interests of those who are identifiable in
the digital phenotype.
4
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
This fails to consider the interests of those who are not identifiable, that is, all other
people who may also be affected when a new digital phenotype is created.
Every time generalizable knowledge is produced by studying a digital phenotype,
there are people who are affected outside those identifiable in the data. E.g. we
discover that smoking causes cancer. This affects all smokers (e.g. higher insurance
prices for all smokers).
N.B. The fact that generalized knowledge harm some people all things considered
may be a reason against producing it; but it may also not be, and in many cases, it is
not, an overriding reason and a reason all things considered for not producing such
knowledge. The benefit for all may outweigh the harm of the few.
N.B.2. The problem of generalized knowledge has been recognized in the past for
genetic data. Many have claimed that genetic data is special because it is shared
among relatives. That is a red herring. If knowledge produced with data from person
A allows inferences about person B, then knowledge production about A creates
potential risks (e.g. privacy and discrimination risks) for person B. That is obviously
the case when you produce genetic knowledge about A and B is A’s twin. But it is
also true in the smoking causes cancer case, even if there are less inferences from A
to B to be made.
2.2 The limits of the libertarian approach
The libertarian approach cares about the interests of the people involved in the
production of the digital phenotype, irrespective of whether or not they are
identifiable. This view has several problems:
oPeople can be involved in data production in two different ways: as persons
causing the data collection process to exist and as persons causing the
particular data to exist. E.g. as a shop owner I tape all my clients (person
causing the data collection process to exist); as a client I am taped by the
shop owner (I am the cause of some of the data collected). The interests of
parties involved in data production may conflict. The libertarian approach
does not provide principles to solve those conflicts (except that people are
free to make individual contracts without use of coercion – violence).
oSome digital phenotypes (e.g. Google searches from the beginning of its
existence) are due to the contribution of millions of persons. Shared
ownership to such phenotypes among the co-creators (e.g. all persons typing
a search string on Google) dilutes the practical value of ownership in terms of
control, for each.
3 Governing the health-relevant digital phenotype
In relation to health we can distinguish three levels of analysis: medical data, data
concerning health (in the GDPR sense) and health-relevant digital phenotypes. The
5
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
health-relevant digital phenotype is the broadest of the three concepts,
encompassing the other two. Tweets about vaccination constitute a health-relevant
digital phenotype (it can be studied for public health purposes) but are not medical
data or data concerning health in the GDPR sense
Case study 1. P4 medicine
Personalized medicine using big data generates a digital phenotype that illustrates
the ethical issues for people who are neither identifiable nor involved in data
production.
The big issue for governance is to protect groups that may suffer harm from
discrimination produced by algorithms that are learned from the data.
It may be exceedingly difficult to identify some of these groups.
It may be difficult to use existing governance models for genomic research which
focus on the protection of a specific population, e.g. endangered ethnic minority,
vulnerable to stereotyping, etc…
Case study 2. Digital epidemiology and the use of internet data in public
health
Another example is the digital phenotype from large platforms such as Google. Here the
issue is that these platforms produce data that may potentially benefit populations, e.g. if
they are used for public health.
But the data collection process and algorithms are not optimized for epidemiology.
Such platforms should at least become more transparent about their data and
algorithms, allowing broadened access to their data and algorithms, in order to enable
researchers (not working for or in direct partnership with Google) to criticize and improve
them.
Conclusions
!
"
0
"
@0
6
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
"
"
/?#
/29
3
#
<
"A
/
"
Reference list:
/5B.737.
B@8C@2
<7&'(=:7
<D4.+3E4
.;45C&&'(=
1FFF###
######F
/5B.737.
BCG@2&'(=
:33!<.4.6
H/..3
5<4.@;
Croatian Medical Journal)-$&*1&'-I(?
@B%5(JJ9Principles of
Biomedical Ethics96E2E7/
@.,5&'(&:5K@.;
Information, Communication & Society()$)*1==&I-J
@/&''-:!@313
57;Journal of Political Philosophy()$(*1&&I9)
5/@(JJ(:257<1
67;
Business Horizons?9$9*1?JI9L
7
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
5#7BH(JL(Cultural Transmission
and Evolution: A Quantitative Approach2121
2E2
5M/(JJ)Self-Ownership, Freedom, and Equality
515E2
.%%45C<NO,/
@/.24&'(-:/
M1.</2
5!;Big Data & Society9$&*1
&')?J)(-(--&=))9
.<(JJJThe Extended Phenotype: The Long Reach of
the Gene&616E2
35.M4&'(9:E@.2
42<;@135
.M4544
/
3M&''J:!!1
3724!
/752@
!;Journal of Medical Internet Research(($(*1((
B&'((:!C2!;
Minds and Machines&($9*1)9JI==
PPP&'(9:6..2M2;
Philosophy & Technology&-$(*1(
PPP&'(=:64.<
2;Philosophy & Technology&J$9*1?'-I?(&
PPP&'(=:M21/.!;!
Group Privacy: New Challenges of Data Technologies
BB@7L?I(''
517
.&'(9:6!..
575<;Journal of
Broadcasting & Electronic Media)L$(*1)JI-)
M %&''LOn Human Rights616E2
4.%&''LWhen Species MeetQ&&9
1E2
4M &''):57<1/3
;The Geneva Papers on Risk and Insurance -
Issues and Practice?'$?*1?L-I9'J
4B%5B8C.2&'():!
@./456H;
BMC Medicine(?19
%3%B&'')Evolution in Four Dimensions:
Genetic, Epigenetic, Behavioral, and Symbolic Variation in the
History of Life51!2
%74@H2%@4%7
@&'():.2;Nature Biotechnology
??$)*19=&I=?
8
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
%>2%.%BR2B
B&''=:7!331
!/3 7B;
Applied Soil Ecology?&$&*1()?I(=9
,B&'(-:E/!77E;BBC News
(&'(-
1FFFF#?J(&='&-
,2&''(Science, Truth, and Democracy6ACS1
6E2
B.<,M,/Q
&'(9:2M1@./;
Science?9?$=(-=*1(&'?I)
B.&'():42.31/5
5;Health Promotion International?'$(*1(-9IL?
PPP&'(=:.573.1!
..I/;Big Data &
Society?$(*1(I)
PPP&'(=:1B.B.BB
7;Leisure Studies(I?
%T2 &'(=:!23
7.;!Tenth International AAAI Conference on
Web and Social Media
1FFFFF!5H7F!5H7(=FFF(
?(=?
@&'(-:.4#<!
132M;InformationL$?*
1FFF&'-L#&9LJFLF?F--
PPP&'(-:34#<!1/
C<;Ethics and Information Technology(I(J
@2/7H
B&'(=:3/1
.;Big Data & Society?$&*1&')?J)(-(==-J=-J
C/7&'((Toward Precision Medicine:
Building a Knowledge Network for Biomedical Research and a New
Taxonomy of DiseaseH.51C/2
C<(J-9Anarchy, State, and UtopiaCS1@
@
6&'():6;!The Onlife Manifesto
B-I(?7!2
2B&'(?:DHM+U21
<M;Bioethics&-$-*1?)-I9('
2@/@&'(?:/7#@/
M<@;Medical Law Review&(
$(*1-(IJ(
Regulation on the Protection of Natural Persons with Regard to the
Processing of Personal Data and on the Free Movement of Such
Data, and Repealing Directive 95/46/EC (General Data Protection
Regulation 2016/ 679)&'(=1FF#F#
F3CFVFW
9
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
X16%BY&'(=((J'('''('(3CMZX6%1B1&'(=1((J1
65
<.%2 &'(9:7B7
@;Science?9=$=&(?*1('=?I('=9
7[B@%@..@%
7@5@35
&'(&:.3;PLOS Comput BiolL$-*1(''&=(=
7[7,&'((:/Q
7671!!
..5;PLoS Comput Biol-$('*1
(''&(JJ
Q3 %.B&'(=:4
.3@.;H46&'(=
1FFFFFFF
Q3 7[B5 %7
@&'():35@.2
4;PLoS Comput Biol(($&*1(''?J'9
H4&'(?The Connected Self: The Ethics and
Governance of the Genetic Individual5E2
H .3B@C&''):75
M122D@+;
Genomics, Society and Policy($?*1LI&(
H .3<CH &''?:5
M@;New England Journal of
Medicine?9J$(&*1((L'I((L9
H/#H/&'(=:H/.
35<4.
@;HH/6&&&'(=
1FFF#F#####
######F
\7MM2<V\4
\@S\&'(9:B5.1
7C.;!Eighth
International AAAI Conference on Weblogs and Social Media
1FFFFF!5H7F!5H7(9FFFL
'9&
10
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
A preview of this full-text is provided by Springer Nature.
Content available from Philosophy & Technology
This content is subject to copyright. Terms and conditions apply.