Consent First! Robots Should Ask Before Recording and Sharing

As part of a larger project to design an Interactive Robot for Ikigai Support (I.R.I.S.), a conversational robot that can assist older adults in recognizing and maintaining their ikigai (sense of meaning and purpose in life). I.R.I.S, a white, 63 cm tall robot, features moving arms and head, facilitating conversations on various topics pivotal to users' ikigai. These discussions include reflecting on a significant photo, contemplating life's past, present, and future, and engaging in open-ended dialogues about what matters most in their lives. Six older adults familiar with our prototype robot from prior participation in our research participated in a 1.5-hour co-design session where we asked about the privacy of recording and sharing with a robot in daily lives. We asked the question, "What does robot privacy mean to older adults?." Below, we summarized the panelist’s feedback on privacy with robots and design considerations for robot designed for older adults.

We began with the questions “What does privacy mean to you?” and “What privacy concerns do you have?” Three panelists expressed concerns about the lack of choice of their privacy in the current technologies. Big Daddy (65/M) noted, “That's pretty much what it was because anybody could listen in...We’re not shielded for privacy from big business listening to us...I don't trust any technology that it's doing what they're saying it's doing. So I give up.” This distrust highlights the challenge of ensuring privacy in an age of pervasive surveillance, named “Big Brothers” by participants: “Even if the robot or the intelligent device tells me it's not recording, tells me it's not saving, or it's not transmitting, I don't trust it.(Big Daddy, 65/M),” and “I think that they're already listening to us and have been for a long time. So you'd have no privacy anymore, you know? (JJ, 70/F).” This could resemble privacy-utility trade-offs, in which individuals tend to be more accepting of technologies processing their personal data, even with some loss of privacy, when the anticipated or realized benefits are high. It also reflects the privacy paradox, where there is mismatch between users and actual behaviors.

Despite their mistrust of the current technology, the panelist, nevertheless, value their autonomy in areas in places where they can control. They expressed preferences for the robot's placement; the Wise One (75/M) said “bedroom and bathroom,” where it is more private, should be avoided, but “family room where that would be good. Like that's like your television room (JJ, 70/F)” or “living room (Anna, 85/F),” where it is more open space, would be good. This feedback suggests that robot design should allow users to control or choose the robot's location, including designs of movable robots.

We then asked the panelists if they had the choice to choose what setting they felt comfortable sharing and with whom they would be comfortable doing so. We have a few `yes’ or `no’ cards that let them choose different scenarios of the person they would be willing to share with. Regarding who to share information with, opinions varied. Three participants were open to sharing with family, depending on the relationship. Discussions on sharing with researchers and healthcare professionals yielded mixed responses, indicating a cautious approach to privacy. JJ (70/F) highlighted the need for consent: “You would have to sign a waiver or an agreement.” The idea of sharing interests for social activities with different robots emerged. The participants agreed on the need for explicit consent, as shared by Middle 52 (70/F): “You're going to go gardening, have to ask. I just thought of [neighbor name]. Would it be okay if I shared with her that you're going gardening?” Consequently, this feedback implies a design requirement for the robot to "ask" for permission before utilizing or disclosing data, ensuring consent is obtained in the initial interaction.

In conclusion, in our panel on creating a conversational robot for ikigai support, six older adults illustrated their views on privacy within the context of robotic assistance. They recognized the constraints of control in corporate settings, echoing notions of privacy-utility trade-offs and the privacy paradox. However, they also emphasized the significance of user autonomy in robot settings, expressing clear preferences for location, information sharing, and the critical need for consent, advocating for a policy of "asking" before any data is shared with others. As we move forward, let's remember the power of asking and the importance of respecting individual choices in the ever-evolving landscape of robot-assisted living.

(Image generated by Dall-E, text modified by Long-Jing Hsu)

This work was modified by the paper submission in the Privacy-Aware Robotics Workshop at the ACM/IEEE International Conference on Human-Robot Interaction 2024.