Kinda Human / Curb

Different aspects of Curb are designed to ensure that the detector integrates itself into an intimate relationship effectively.

• The Visual Feedback DisplayedSpecial attention was paid to the visual feedback of Curb, specifically the aspect of feedback resembling a heartbeat—slow and regular when listening, and fast when an indiscretion is detected. The color of the LED lights also change when an indiscretion is detected to depict an activated state. In addition, I decided to scatter the lights throughout the detector to bring life to the piece in its entirety rather than just a small part of the detector (i.e., the four lights in the center of Google Home; the ring around an Alexa).

When triggered, LED lights beat faster and change colors

• The Look and Feel of the ObjectI made an explicit decision to give Curb a form that was very different from what a user would expect from a smart speaker today, in order to combat availability bias. Curb was designed to have a more natural shape, resembling a kidney. Instead of filtering blood, Curb filters language. By making references to a natural form, the prototype more easily integrates itself into a couple's home environment.

Curb on a wall

• The Language it would employI avoided very forceful language, in order to create a collaborative environment that put the onus on the individual to participate in some form of self-introspection. As a result I designed the detector to ask the partner, Is that what you meant to say? This conversational approach opens a line of dialog which conveys a degree of uncertainty but also serves as a prompt for reflection.

To test the concept, I created a video. The video introduces the detector and portrays a sequence of four different couples encountering some form of a breakdown in a conversation. After every breakdown, Curb interjects by asking the partner, who was the instigator of the breakdown, if they meant to say what they just said.

Sample dialog from Bedtime Visions’s concept video

The prototype intentionally includes a variety of voices to represent the diversity of intimate partners and a range of different shots to depict the potential versatility of the detector (i.e., the detector could be place on a wall, ceiling, or table surface in a family room, kitchen, or bedroom).

Curb's user sentiment was less favorable than that of Intimately, but still garnered more positivity than negativity. While some participants saw it as annoying, stupid, and overbearing, others saw it as neutral and empowering. Those who saw it in a negative light said using Intimately could represent "admitting that one’s relationship is bad." Whereas those that viewed it in a positive light believed it "could save a marriage" or "be helpful to people having a hard time talking to their spouse." Other insights gathered from research participants include:

• Curb could evoke a confirmation bias if a couple is given the opportunity to select specific measures for detection.Similar to Intimately, if Curb provides a user with too much power in personalizing the detector, opportunities for enhancing an intimate partner's capacity for expression and understanding could be missed.
• Curb could confront challenges affecting behavior change.Similar to Intimately, Curb may struggle to effect behavior change on a consistent basis if the same communication breakdown activates Curb time and time again. To enhance a partner's capacity for expression, it may be beneficial to explore different forms of support Curb can provide.

Curb's design was inspired by smart speakers and designed to explore the different ways to convey information to intimate partners. I sought to gain a better understanding of the ways such an intervention could integrate into the environments that couples inhabit. The following is a collection of principles I gathered while testing Curb.

• Curb should support a diversity of conversations and contexts.Like Intimately, Curb should deal with conversations beyond those of conflict. For instance, Curb could celebrate a conversation that builds understanding between a couple, instead of only bringing attention to breakdowns. In turn, Curb could increase people's appreciation and perceived relational value for themselves.
• Curb should allow for multiple forms of adjustment.Curb, like Intimately, should allow room for a user to adjust the parameters and patterns the detector is seeking, resulting in an open channel back and forth between the couple and Curb; essential for conversation symbiosis.
• Curb should provide a user the opportunity to dismiss a detected indiscretion.Similar to Intimately, Curb should provide an environment for people to reflect and evaluate the specific exchanges that trigger Curb. As a result, a couple could avoid situations where Curb is activated time and time again in response to exchanges that the couple has deemed as positive.
• Curb should afford numerous opportunities for reflection.A user should be given a variety of opportunities to reflect on the moments Curb is triggered, ultimately giving users time to acknowledge and consider instances of breakdowns and wins in their communication with their partner.
• Curb should evolve its voice based on a couple and their communication.The specific language and delivery used by Curb could change based on the couple and the situation in which they find themselves. It is naive to think that every couple or even the majority of couples can be successfully reached through the use of the same language and delivery of that language. Thus, a diversity of language is needed to reach a large portion of couples.

Storyboards based on Curb that depict potentially unfavorable directions focused on the following lines of questions.

• Curb being biased towards certain forms of expression.
• Curb normalizing dysfunctional behavior over time.

A number of research participants saw value in Curb, but also stated that after using the tool a few times they would be annoyed and possibly agitated by having this virtual-assistant voice interrupting their conversation. This caused me to consider the feasibility and benefit of having a similar dynamic, but posture the interruption in a less personal, calculable, and predictable way. With this in mind, I developed an iteration of Curb, where instead of having the agent ask the partner "Is that what you meant to say?", Curb would play a popular piece of music with lyrics related to that conversation. The intention would be to surprise an individual and enable them to take a reflective approach towards their conversation with less of the annoyance and agitation associated with the virtual assistant voice.