IEEE Technology and Society Magazine - December 2019 - 52

specifications. Additional stratification could be introduced to integrate data resulting from the Yamagishi &
Yamagishi's General Trust Scale in order to measure
trust in different contexts: trust in general, interhuman
trust, trust in robot developers, trust in robot implementers, and to compare the results with trust in machines.
Moreover, this assumption is in line with the intentions of the already mentioned recent European Commission Guidelines for realizing trustworthy AI: ensuring
that these advanced systems are human-centric and
grounded in ethical principles and values. For example,
operating transparently is considered a key issue for
building and maintaining people's trust both in developers and implementers of advanced robotics and robotics themselves [22].
As for trust in robots in general, surveyed in Section
B of the questionnaire, a more positive trend compared
to what emerged in Section A was found. Results indicate that a large part of the sample regarded these systems very favorably (79.4%), consistent with a high
percentage of interest in science and technology (87%),
and despite concerns about the possibility that robots
could steal jobs in the future. These results suggest that
the propensity to be trustful toward advanced robotics
could increase in the future, along with the increasing
use of technology among young people. However, effective interventions could be implemented to reduce the
threat to confidence in robots represented by the real
concern (68% of our sample) that new technology can
take away jobs from humans. Future research is necessary aimed at defining existing and new jobs consistent
with Industry 4.0, such as those that are very tiring,
menial, or dangerous for humans.
The present survey also identified a significant gender difference in trust in robots: males trusting more
than females. This data, which supports previous studies [28]-[30], invites further research to investigate
female attitudes and interactions with robotics so that
new machines can be designed taking into greater
account gender differences among users.
Regarding possible future scenarios of advanced
robotics, respondents feel comfortable with machines
only in 6 future scenarios of advanced robotics out of the
14 presented in the questionnaire. Acceptance of robotic
machines seems to depend on the nature of the role they
perform in the 6 most desired application areas.
Respondents were less likely to desire robots as
caregivers for their own grandparent or as teachers;
roles in which the relational and human sharing aspects
are clearly very important. In contrast, a high percentage of positive responses to questions about robots in
mostly technical roles was registered. Thus, it can be
deduced that respondents' preferences correspond with
the idea of robots as a means to reach an end.

Interestingly this connection is inconsistent with
respect to robots in positions of command, which
obtained a very high refusal rate (90%). Although the
response suggests the need to rephrase the question to
evaluate situational trust better [28] by specifying the
context(s) in which the robot would exercise the role of
command, its potential usefulness in indicating paths for
future investigation can be seen. Specifically, when situations switch from those in which robots are under human
control, to situations in which they control humans or
their matters of life and death (city safety, care for a family member), the degree of confidence toward robotic
machines fell significantly, until it became almost nil. The
data here appears in accordance with the low level of
general trust among the sample discussed above: people
who are wary toward other humans are also wary toward
robots, but to a greater degree when it comes to entrusting something important to them.
Consistent with these observations, trust among
younger generations needs to be improved. First as a
foundation for building confidence toward nontechnical
application areas in which robots could assist and
relieve humans in their work. Second, for building and
maintaining people's desire for comfortable robots on
firmer ground by further supporting the same tendency
shown in previous studies [20], especially in those scenarios in which the presence of robots is widely desired,
and not simply a function of supporting human work.
The tendency to anthropomorphize advanced robots
was also investigated, to assess how this may influence
or be connected to trust in robots.
Specifically, respondents were uncomfortable with an
anthropomorphic appearance for 5 out of the 6 most
desired application areas, including robotic machines as
house cleaners, street cleaners, taxi drivers, soldiers, and
cooks. Several factors that may have contributed to this
nonpositive attitude toward anthropomorphism have been
identified. For example, respondents preferred that robots
carry out roles that are of a tiring, menial, or dangerous
nature. This probably explains the nonpositive tendency
toward anthropomorphic projections in robots. In other
scenarios, the role did not require a high level of interaction
between humans and nonhuman agents and therefore, in
the absence of relations to be facilitated, did not move the
respondents to desire anthropomorphized robots, in line
with what has emerged from a recent analysis [31].
Only in one case of technical robots, namely that of
personal trainers, was there a preference for anthropomorphism (56%), suggesting the usefulness of self-projection in robot design where learning through imitation
is concerned.
Anthropomorphism was also preferred in the other
scenarios where, regarding friends, the roles were
considered relational and affective, despite receiving

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

∕

DECEMBER 2019

IEEE Technology and Society Magazine - December 2019

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - December 2019