Fair incentives for covid 19 tracking

Can one fairly incentivize the
adoption of a digital contact
tracing app?
Michele Loi
Senior Researcher
Institute of Biomedical Ethics and the History of Medicine
Digital Society Initiative
University of Zurich

Spoiler:
The answer is yes, but only with “enablers that are not incentives” (in a
narrow sense of the term, to be explained) and still greatly facilitate the
adoption (so, it is an incentive, in a broader sense).

Contents
• Section 1: The ethical problem with incentives
• Section 2: Proposed solution
• Section 3: Answering the data quality, corruption, informed consent
objections
• Section 4: Proposed solution to the fairness objection
• Section 5: Is a mandatory app preferable?
• Section 6: Information gaps

Definitions
1. Digital contact tracing apps: https://en.wikipedia.org/wiki/COVID-19_apps
2. Incentives: I use two possible definitions. The proposed solution is an
incentive in sense 1 but not in sense 2.
P= agent; F= action
Incentive 1: x is an incentive for P to F if x is a condition outside P and P is
more likely to F if x occurs
Incentive 2: x is an incentive for P to F if P triggers motivations to F that P
would not have in the absence of x

Section 1. The ethical problem with incentives
Non – academic COVID 19 emergency relevant sources:
Luciano Floridi’s reflection on digital contact tracing apps
https://thephilosophyofinformation.blogspot.com/2020/04/mind-app-
considerations-on-ethical.html
Informal audition of the President of the Italian Privacy and Data
Protection Authority (Garante per la Protezione dei Dati Personali),
Antonello Soro
https://www.gpdp.it/web/guest/home/docweb/-/docweb-
display/docweb/9308774

Incentive type distinction. If I understand Floridi’s terminology (see
Floridi, cit.)
• Endogenous = advantages connected to app outputs and use
(advantage from the app)
• Exogeneous = advantages connected to merely having/installing the
app (advantage for [getting] the app, in Floridi’s words)
Both appear hugely problematic.

• Endogenous = advantages connected to app outputs
• Data quality problem: if you receive a benefit only when the app says you are
not a contact of an infected person, you have an incentive to trick the system
to avoid proper recording of your status (see Floridi, cit.)
• Informed consent problem: if there are special rights and privileges
conditional of an app outcome, consent for the app cannot be considered
voluntary (at least according to Italian Data Protection Authority, cit.). E.g.
Chinese system: have to show your app status to do X, Y in society: cannot be
considered voluntary, informed consent does not obtain.

• Exogeneous = advantages connected to merely having/installing the app
• E.g. money. Problems
• 1) data quality problem (e.g. Titmuss, gift relationship): people who do it for the
sake of money, not public health, are less reliable in their app use
• 2) corruption/ethics crowding out argument (Titmuss, Deci, Frey, and many
others): sends the wrong signal to the morally motivated part of the population,
may be a counter-incentive for ethically motivated people to participate (lowers
adoption rate of “careful” users: risk of alienating most potential users + lower
average data quality)
• 3) fairness problem: as ownership of advanced smartphones (necessary for
implementing the protocols) tracks social inequality, any benefit provided to
users is a gift to those who are already advantaged; as the service has little
positive externalities when adoption is limited (e.g. to users in urban wealthy
areas, etc) “benefits the worst – off” argument not persuasive. (Floridi, cit.)

• Possible solution: move to compulsory system, so you do not need to
rely on incentives -> has many problems, discussed later (section 6).
• N.B. Possibly not even a solution, all but the most coercive forms of
enforcement of a compulsory system also raise data
quality/corruption issues.

Prima facie conclusion:
1) The only motivation on which it is ethically unproblematic to rely is
ethical motivation, i.e. acting for the sake of public health goals,
wanting to prevent the spread of the disease.
2) No incentive is ethically viable.

Conclusion:
1) The only motivation on which it is ethically unproblematic to rely is ethical motivation,
i.e. acting for the sake of public health goals, wanting to prevent the spread of the
disease.
2) No incentive is ethically viable.
Strategy to reject conclusion:
Reject (1): some prudential (partly self-interested) motivation (protecting the health of
one’s contacts and taking timely measures for protecting one’s own health) is also ethically
unproblematic in this context
Reject (2): incentives that merely remove obstacles to action motivated by self-interested
and ethical motivations to prevent contagion and protect personal health are ethically
viable.
This will be argued in section 2.

Section 2. Proposed solution
• The solution has two parts:
1) an “incentive which is not an incentive”* tackles the data quality,
corruption, and informed consent problems
2) a specific economic subsidy for smartphones + free minimal
functional device for digital contact tracing for non smartphone
users tackles the fairness problem
*it is or isn’t on two distinct definitions of “incentive”

Section 3. Proposed solution
The proposed solution:
Legal guarantee of rapid clinical testing +
[A legal guarantee that the contacts traced have timely access to clinical COVID 19 tests, to
ascertain their actual COVID 19 positive/negative status]
+ a context in which COVID 19 testing will be scarce +
[assuming: scarcity is not artificial and strategic: it results in spite of the best efforts of
society to secure testing to as many people as possible, under reasonable resource
constraints]
+ compensation of the relevant digital divide (see fairness)
[economic subsidies to buy smartphones, conditional on electronic proof of successful
installation of the app (which must be easy to immediately remove) to mitigate cognitive
and economic disadvantage in app adoption (efficiently) + free simpler Bluetooth enabled
device for non smartphone users (when digital divide not due to smartphone cost +
installation challenge)]

Section 3. Proposed solution: motivation to
install/use the app
Why a legal guarantee of rapid clinical testing will facilitate the adoption of COVID 19
digital tracking apps:
1) People intrinsically strongly desire knowledge of their infected and infectious status to
protect their own and other people’s health
2) People want to assess their risk of having being infected, and if at risk, they strongly
want to have a access to a test (for health reasons and for preventing harm to friends
and relatives at least)
3) People understand that, if they use digital contact tracing, it is easier for them to know
if they are at high risk of infection and if they are actually infected and infectious
And
If, as foreseeable, in most EU countries and the USA COVID 19 testing will be scarce
(assumption of natural scarcity, not artificial/intended one), the motivation in question will
not be easily satisfied in general.

Section 3: Proposed solution/what is the
difference morally
An incentive (def 1) which isn’t an incentive (def 2)
Remember the definition of incentive?
P= a person F=an action
Incentive 1: x is an incentive for P to F if x is a condition outside P and P
is more likely to F if x occurs
Incentive 2: x is an incentive for P to F if P triggers a motivation to F that
P would not have in the absence of x

Section 3: Proposed solution (analogy
argument)
An incentive (def 1) which isn’t an incentive (def 2) may differ morally from an incentive (def 1) that
is also an incentive (def 2):
Example:
The government wants to augment the proportion of students in STEM subject, it considers 4
different incentives:
- - Strategy 1: people with a STEM degree get priority treatment in otherwise long administrative
procedures (endogenous, is incentive according to def. 1 and def. 2)
- Strategy 2: the government pays people to get STEM degrees (exogeneous, is incentive according
to def 1 and def 2)
- Strategy 3: the government creates more favourable circumstances for high-end technology
companies, increasing demand for STEM degrees, that leads to higher salaries and lower
unemployment (endogenous, is incentive according to def. 1 but not def. 2)
- Strategy 4: the government improves science teaching at all educational levels (exogeneous, is
incentive according to def. 1 but not def. 2)

Section 2. Proposed solution. (Analogy
argument)
Moral difference between strategies (1 – 2) vs. (3 – 4)
(1 – 2) confer privileges to a specific category of citizens so that they study
STEM subjects
vs.
(3 – 4) citizens are more likely to study STEM subjects as a result of implementing
policies that fulfil the general needs of all citizens (considering social justice)*
*Assume for argument’s sake that facilitating the creation and maintenance of
technologically advanced firms is justified by economic considerations, taking
social justice in consideration (it is reasonably expected to boost the economy and
improve everyone’s condition, including that of the worst off). Assume improving
science teaching in elementary schools is justified by both considerations of
intrinsic or broad and instrumental value of education and by the economic
considerations above.

Section 2. Proposed solution.(Analogy
argument)
Psychological difference between strategies (1 – 2) vs. (3 – 4)
In (1 – 2) the incentive creates new motivations for people to
pursue STEM subjects at the university (administrative advantages,
money)
vs.
In (3 – 4) the incentive makes it easier for citizens to pursue STEM
subjects at the university based on motivations that citizens have
anyway independently of the incentive
Assume: getting a remunerative job and/or getting a STEM are
motivations that the beneficiaries of the incentive have anyway.

Section 2. Proposed solution.(Analogy
argument)
An analogous argument can be made in favour of incentives that are
similar to 3 and 4, but differ from 1 and 2, morally and psychologically.
It can be argued that our proposal (slide 13):
a) citizens are more likely to install and use the app as a result of
implementing policies that fulfil the general needs of the citizens
and are independently justified (i.e. ensuring rapid testing to
citizens at higher than average risk)
b) citizens are more likely to install and use the app driven by the
same motivation that they would have in the absence of the
incentive

Section 3. Answering data quality, corruption,
informed consent objections
1) Data quality objection:
(Endogenous) Legally securing rapid access to clinical test for people who are at
high risk of contagion does not compromise data quality, because the
advantage (getting a clinical test) is not valuable intrinsically, and the only way
to get access to the advantage via the app is by increasing one’s chances of
infection (an intrinsic harm). Nor is that a problem for the smartphone subsidy:
those who do not like the app can immediately uninstall it.
(Exogenous) Data quality is not affected by the quality (selfish vs. moral) of
motivations. See response to “corruption of motivations” objection for details.

How the solution avoids objections:
2) Corruption – moral crowding out objection:
Securing legal access to clinical test for people who are at high risk of contagion
does not corrupt the motivation for using a digital contact tracing app. The
motivations for wanting immediate access to a clinical test and the motivation
for using a digital contact tracing app are essentially the same:
- to protect other people from contagion
- to take timely appropriate measures to protect one’s own health
Nor is does the smartphone subsidy: those who do not like the app can
immediately easily uninstall it. Only people with the right motivation end up
using the app.

How the solution avoids objections:
3) Informed consent objection:
Securing legal access to clinical test for people who are at high risk of contagion is compatible with
informed consent being voluntary on the strictest of criteria for voluntariness for using the app, as
long not special right and privilege is attached to using an app. This can be achieved if the legal
guarantee that the contacts traced have timely access to clinical tests to ascertain their actual COVID
19 positive/negative status obtains by virtue of:
a) a general legal guarantee that confers priority in access to clinical tests to people at special risk
b) a specific legal provision that includes people singled out by the app as in a possible chain of
contagion as people at special risks in the scope of (a)
c) the fact that a digital contact tracing app is not the only way to certify the qualification of person
at special risk in the scope the law (a). This includes at least contacts discovered with traditional
contact tracing (which must be in place) and other at risk categories specified in such law.
d) a legislative framework that in fact enables and favours other digital risk-assessment tools and
sets in place a validation mechanism (e.g. in terms of accuracy, etc) for inclusion of further tools
(i.e. delivered by the free market) to assess the condition of “being at special risk”
NB: (c) and (d) are meant to ensure that effectively, priority testing is not conditional on having the
app (it is not just a legal fiction).

Section 4. Proposed solution to the fairness
problem
The fairness problem:
1. ownership of advanced smartphones (necessary for implementing
the protocols) tracks social inequality,
2. any benefit provided to users is a gift to those who are already
advantaged;
3. the service has little positive externalities when adoption is limited
(e.g. to urban wealthy areas, etc) “benefits the worst – off”
argument not persuasive. (Floridi, op cit)

problem
Proposed solution
1) The government produces and distributes for free Bluetooth-based
basic digital tracking devices on which a basic version of the digital
contact tracing protocol can be run, optimized for the privacy-
protecting protocol of the apps, and requiring minimal digital skills
to operate (basically: reading the informed consent form and
clicking on ”yes”; de-activation is guaranteed by intentionally failing
to charge batteries)
2) The state subsidizes the purchase of smartphones able to run the
app by providing discounts equal to the cost of delivering free
Bluetooth design

problem
Proposed solution
Reasons why (1) is essential:
- unfairness is created not only by unequal opportunities for people
who cannot afford a smart phone to access clinical testing, it is also
created by unequal opportunities for people who object to using a
smartphone to to privacy measures (e.g. special privacy vulnerability)
and a digital divide of non economic nature.

problem
Proposed solution
Reasons for (2):
- (a) Providing a discount for the purchase of a new phone can be hugely more
efficient, as such phone will generate additional utility (better access to the
internet, consumerist pleasure, status)
- (b) People who consider purchasing a more advanced phone also for the sake of
its ability to run the app should be put in a condition to do so; it would be unfair
for them not to benefit from the public subsidies to the same extent as people
who dislike or cannot use smartphones,
- (c) even if the subsidy serves other types of motivations (general internet access,
consumerist pleasure, status) meeting some of these needs (i.e. better access to
the internet) is also required by fairness because it mitigates the digital divide,
which is a cause of other forms of unfairness (increasingly so under a Covid 19
emergency)

problem
Proposed solution
Details about (2):
a) To ensure that the purchased phone is able to run the digital contact tracing app, the client will only
receive the subsidy after sending an electronic proof of the download and successful installation of the
app.
b) The phones do not come with pre-installation, the discount is authorized after installing, and is valid
irrespective of uninstallation. The beneficiary must install the app her/himself (opt in, not opt out), but
of course may ask the help of shopkeepers. After this is completed, the procedure to authorize the
discount is activated, and the client is shown a clear message, e.g. “now you can keep the app on and
contribute to digital contact tracing, and you are also free to uninstall it without losing your discount. We
encourage you to keep the app but you are not legally obliged to do so.”
c) The discount on the new phone is not meant as an economic incentive to use the app, but closing a
resource gap to a tool everyone needs
d) The required installation of the app (with permission to uninstall) is not meant as a nudge (i.e. opt out vs.
opt in). It is only meant to address the cognitive and time resource gap, ensuring that people are more
likely to receive assistance to install the app (customers or new phones are more likely to be helped by
shop keepers)

Section 5. Is a mandatory app preferable to a
mandatory one
• Fair incentive problem only exists if it is unethical to make the app
compulsory. Is a compulsory app unethical?

mandatory one
• Non feasibility of enforcement of a mandatory app:
1) Enforcement of legal requirements to use / show the app (for those
who have a viable phone) in public life
Fairness objection: this type of enforcement would be unfairly
discriminatory in the absence of the “fair solution”- it only limits the
freedoms of wealthy users who can afford expensive smartphone. This
objection is avoided by the same solution offered for the incentive
approach.

mandatory one
• Non feasibility of enforcement of a mandatory app:
1) Enforcement of legal requirements to use / show the app (for those who have a
viable phone) or free device in public life
• Data quality objection: short of securing the physical attachment of the device to
the body, people who do not agree with the purposes of the app or fear negative
externalities will act in ways that reduces data quality (e.g. “forgetting” the
telephone at home)
• Non-proportionality perception objection: if the Bluetooth device is securely
attached to the body (like an ankle bracelet from prison) this is widely be
perceived as disproportionate coercion. Whether the coercion is in fact
disproportionate or not (I think it is, but argument takes space), backlash against
such measure is very probable (if only for its symbolic implications) and no
democratic government will be able to push this through.

Section 6. My information gaps (other gaps
may not be as obvious to me):
1) Economic/epidemiological: how likely are clinical tests to be scarce?
2) Economic/epidemiological: how advanced must clinical test capacity be, in
order to fulfil the requirement of timely access to clinical tests for all the
contacts traced?
3) Psychological: how likely is securing access to clinical testing + removal of
economic disadvantage in access to suitable smartphone technology to
facilitate app use?
4) Viability and possible cost of the free simple digital-contact-tracing optimized
device (slide 22)
5) Legal: does the proposal involve a conditionality to rights/privileges which
compromises informed consent as a basis of legal processing of the data?
6) Legal: in the argument against mandatory digital contact tracing apps: can one
say that a legal right is violated by a ankle-bracelet-like device (with legal
punishment for non carriers)? What would courts say about proportionality?

Fair incentives for covid 19 tracking

Recommended

Recommended

More Related Content

Similar to Fair incentives for covid 19 tracking

Similar to Fair incentives for covid 19 tracking (20)

Recently uploaded

Recently uploaded (20)

Fair incentives for covid 19 tracking