IEEE Robotics & Automation Magazine - June 2019 - 65

connection with people" or the "robot creating and sharing
items." The mid-openness participants (n = 4) were more
open to "the robot creating and sharing items" compared to
the low-openness cluster. The high-openness group (n = 5)
was the most embracing of the robotic functions, with the
desire for them rising above 50% in all categories. After the
study, the low-openness group (n = 4) became more polarized regarding all functions. The mid-openness participants
(n = 3) became more open toward all functions, except
"connecting with people." The high-openness participants
(n = 6) were more accepting of the "robot creating and
sharing items" and "suggestions," were less accepting of
"connecting with people," and were comparable on
"reminders and alerts" and "information." The mid-openness group shifted toward greater openness to the technology, while the low-openness group moved toward being
more skeptical of the technology. The high-openness group
shifted slightly while maintaining a greater than 50% desire
for all functions.
A 2 # 3 Fisher exact test determined that there was a significant difference in overall social robot preferences among
the three clusters and between the initial and final sessions
( p 1 .0001))))) . Each individual category was found to be significant as well:
1) mediating connections with people: p 1 .0001))))
2) information: p 1 .0001))))
3) reminders and alerts: p 1 .0001))))
4) robot creates and shares items: p 1 .0001))))
5) suggestions: p 1 .0001)))) .
Qualitative Results
Participants' opinions throughout the design kit session provided valuable insights into older adults' reasoning for social
robot function preferences. Our analysis revealed 16 themes
that were broadly grouped into five categories:
● technology concerns
● relationship to technology
● limitations of technology and use
● imagining in daily life
● technology features.
A full list of themes related to each category is provided in
Table 2. Technology concerns involved such themes as concern for security and privacy; fear; and a collective worry for
society. Participants often described their relationship with
technology, discussing user confidence and autonomy over
technology. The limitations of technology included operation difficulty, expectation gaps, and inability to conceptualize potential agent actions because of current technology
inexperience. When imagining the robot in their daily life,
participants discussed the logistics of how it would function
and their design recommendations for the robot. The technology features discussed included the role of the agent, novelty, portability, and personality.
The themes most mentioned by the participants also differed based on their openness level (low, mid, and high)
identified using the design kit (Table 3). Qualitative data

were analyzed through the proportion of participants who
discussed the theme. Between the initial and final interviews,
technology concerns increased for all clusters, driven by the
low-openness participants discussing it the most (50%). Participant clusters varied with regard to their relationship with
technology. The low-openness participants decreased, while
high- and mid-openness residents remained the same. The
low- and mid-openness clusters discussed the limitations of
the technology the most. All clusters discussed imagining the
robot in their life and technology features more in the final
session. A Fisher's exact test analysis between the initial and
final interviews for the five categories showed a significant
trend difference (p 1 .05) ) .
These results highlight specific themes that changed
between initial and final interviews, including security and
privacy, fear of the agent, collective worry for society, inability to conceptualize the social robot in their life, and the
design of the social robot. With regard to security and privacy, participants discussed it less during the beginning of the
study (25%). More participants discussed this during the
final interview (58%), commenting on security, autonomy,
and adoption concerns. Participants mentioned, for example, "privacy concerns, because it's another intrusion into
[their] lives. It's artificial and it's technological and [they]
don't trust them in technology" (P19). During the initial session, some participants mentioned their fear of the technology having too much autonomy (13%), emphasizing "it
should be a person, not a robot." (P7). Fear was also prevalent in the final discussions (33%), a majority (50%) from
the low-openness group. P11 echoed P7's concern from the

Table 2. The themes for the categories.
Category

Theme

Technology concerns

Security and privacy
Fear of technology
Collective worry for society

Relationship to technology

Autonomy over technology
User confidence
Personalization
Expectation gaps

Limitations of technology
and use

Difficulty operating technology
Inability to conceptualize
agent interactions
Annoyance of technology
Imagining in daily life

Logistics of functioning
Design recommendations

Technology features

Role of the agent
Personality of the agent
Novelty of the technology
Portability of the technology

JUNE 2019

IEEE ROBOTICS & AUTOMATION MAGAZINE

IEEE Robotics & Automation Magazine - June 2019

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - June 2019