IEEE Technology and Society Magazine - June 2021 - 45

devices and apps more generally, and this will have
been exacerbated during the pandemic with people
spending more time online. You can liken this to the
recovery from " Long COVID, " where people have
long-term symptoms. We are going to have to deal
with this problem too, in particular, protecting vulnerable
groups, among them the younger generation
which has moved their lives almost entirely online,
interacting through screens with their formation of
social relationships disrupted and having their education
and employment prospects disrupted. These
people will need sensible, constructive solutions,
and support.
SW: I would like to wrap up with a word and
theme which both of you have powerfully highlighted:
trust. How do you foresee technology during
this pandemic affecting and transforming substantive
qualitative human values such as trust?
JP: There are many initiatives that can be looked at
in this regard [3]. For instance, there's the concept and
practice of design contractualism, where engineers
make ethical choices and encode them in the system,
and their ethical decisions are then made visible in the
interface. This harks back to Tim Berners-Lee's initiative
in linking data to give people far more control and
ownership over their own data. I have seen mention
of a kind of Hippocratic Oath for ICT practitioners and
also a very interesting article by Luccioni and Bengio
(a winner of the ACM Turing Award), on making moral
choices about the use of deep learning and applications
of artificial intelligence [4].
Then there's the whole issue of informed and meaningful
consent-the pretence that when you make one
of these contracts you have actively read and understood
the terms and conditions. We can turn this on
its head to say: you know these are my rights and you
have to respect them, not just for privacy but also for
attention, and for the data itself, and for knowledge
not only of who is stipulating something but where
do they get it from. My own technological solution to
this is called PlatformOcean [5], where we're trying to
empower communities and make sure they can keep
all their own data within the platform.
Then there is an exploitation of the automation
bias, and this is distinctly problematic when you get
complex situations, such as when people start delegating
cognitive skills to the machine. We see this with
voice-activated video assistance. The generation growing
up with these starts " outsourcing " to Siri or Alexa,
rather than relying on their own mental capacities.
June 2021
This might be okay when the question is " Siri, what's
4 + 4? " But it starts becoming more problematic when
your question is " Should I get vaccinated? " or " Who
should I vote for? " This exploits the automation bias
in a way that can undermine an individual's decision-making
skills and can become manipulative.
MT: Artificial intelligence and digital technologies,
more broadly, have great potential to improve
our lives. Indeed, they are already improving our
lives: we can work, we can get educated, we can
talk, we can develop science much better than we
could have without digital technologies. Having
said this, these technologies are coupled with serious
ethical risks: lack of transparency and accountability,
and disparities in access to and the ability
to control these technologies are good examples of
these risks. Risks also come from the over reliance
on AI and digital technologies. This is problematic,
insofar as AI-based or digital solutions are not necessarily
the optimal ones in all circumstances, and
they always pose important ethical challenges. So,
in some cases, relying uncritically on these technologies
June have a negative impact on rights and
values, while not leading to an effective solution.
My second point is about trust as a concept in
itself. Trust is an easy word that refers to a very complex
concept that philosophers, psychologists, ethicists,
economists, and cyber-security experts consider
in divergent ways. But there is a common element.
At a high level, trust is a way of delegating a task and
choosing not to control how the task is performed. The
lack of control is important here; it is what makes trust
a facilitator of social interactions. For instance, I don't
control how a doctor makes a diagnosis about his or
her patients; I delegate the task, in that way saving time
and energy to do my own job. And you can imagine
this occurring for very many functions in society.
There is a risk though, that with this relationship
between the lack of control and delegation if there
is too much trust in a society, we lose accountability
because you lose any chances to perform control.
Yet, if there is too little trust, the system doesn't work
because you have to check everything. So, there is a
balance to be struck.
There is a further element to trust, which is realizing
what is the trustworthiness of the trustee. This is
your assessment of the agents in whom you decide
to place your trust. Some of the literature merges this
with the concept of reputation: the idea that someone
has been able to perform that task to date, so they will
45
IEEE Technology and Society Magazine - June 2021

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - June 2021
