IEEE Technology and Society Magazine - September 2022 - 45

boundaries between institutionalized and folkmedicine
are not discrete [5]. For instance, Meta has
recently been sued following a discovery that its
Pixel app, used for scheduling by many hospitals,
was collecting patient information and sending it
to Meta [65]. Thus, while users had the reasonable
expectation that scheduling information would be
shared with the hospital, they would not reasonably
expect that sensitive information (e.g., appointments,
prescriptions, and medical conditions)
would be provided to Meta and could be aggregated
with other information.
Policies with privacy implications (e.g., EULAs,
privacy agreements) can be opaque and excessively
lengthy. The probability that users will read their
contents is low. Consequently, such consent cannot
be considered valid. In that smartphones represent
a domesticated form of technology that might be
associated with high trust due to a user's familiarity,
users might be unaware of the nature of the biometric
information (e.g., fingerprints and faceprints)
or behavioral information (e.g., sleeping and exercise
patterns) collected by folkmedical applications.
Indeed, such information could be used to make
uncanny predictions to which the user might not
consent were they aware of the nature and extent of
the data collection.
The quantity and nature of data that is collected
by these applications can lead to significant cybersecurity
issues. Unlike other data breaches (e.g.,
exposure of passwords, social insurance/security
numbers) which pose a risk but can eventually be
mitigated, a user's health records contain information
that is intrinsic to the data subject (e.g., personal/
medical history, genetic information, and fingerprints).
In this case, a data breach can have significant
social consequences. Moreover, these datasets
can be vast. For instance, direct-to-customer genetic
services such as 23andMe have collected data from
over 1,000,000 consumers [66]. On its website,12 the
company reported eight data requests (pertaining
to 11 users) in its Transparency Report, all received
from the United States. Crucially, data protection is
limited in this jurisdiction. As the company website
notes " ...the Genetic Information Nondiscrimination
Act ('GINA') ... protection against discrimination
by employers and health insurance companies for
employment and coverage issues, respectively, has
12Updated, January 12, 2022, from https://www.23andme.com/en-ca/transparencyreport/
not
been clearly established. In addition, GINA does
not cover life, long-term care, or disability insurance
providers. " Thus, large datasets derived from folkmedical
technologies have an ambiguous status that
could be exploited by these organizations or third
parties without users being aware of the nature,
extent, and accessibility of their data.
Changes in information and treatment seeking.
Folkmedical technologies can act as a gateway or
supplement to traditional institutionalized health
services. Such technologies can also provide access
to historically disadvantaged or underserved communities
that lack the requisite financial resources
or geographic proximity that would allow them
to receive quality healthcare. However, we must
also acknowledge that if users cannot differentiate
between folkmedical knowledge provided on the
Internet and social media and that provided by
HCPs and expert systems, they might fail to seek
out treatment at a critical juncture in the course of
an illness.
Over time, folkmedical healthcare technologies
might lead to functional creep, such that users prefer
the use of this technology relative to HCPs for primary
care rather than as a supplement depending on context-specific
factors. For instance, while clinicians
might be preferred in many situations (e.g., cancer
and neonatology), nonhuman agents might be preferred
to humans, such as when the health condition
is believed to be minor (e.g., monitoring sleep or
blood pressure) or when there is a social stigma or
legal issues associated with the condition (e.g., sexually
transmitted or IV drug-related infections). Ideally,
A/ISs should be designed to acknowledge both their
limitations and their embeddedness within a SSH, for
example, directing the users to HCPs. Thus, designers,
distributors, and HCPs must ensure that they have an
accurate understanding of the interactional patterns
between users, folkmedical technologies, and institutionalized
healthcare services.
" Post-Truth " era healthcare and expertise conflict.
In institutionalized healthcare, the increasing accuracy
of medical A/ISs leads to the possibility that A/
ISs should replace HCPs in certain EPAs due to comparative
reductions in cost and time or increases in
reliability and accuracy. The most likely scenario
will be some form of integration within institutionalized
healthcare [4]. Crucially, while A/ISs might
have comparable abilities, if they are not formally
validated and integrated within institutionalized
September 2022
45
https://www.23andme.com/en-ca/transparency-report/ https://www.23andme.com/en-ca/transparency-report/
IEEE Technology and Society Magazine - September 2022

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - September 2022
