Kinda Human / Bedtime Visions

Multiple features of Bedtime Visions were designed so that couples could uncover and develop visions of their future together.

• The Context of the ExperienceDesigned specifically for couples at bedtime, Bedtime Visions is intended to create an environment in which partners are able to lower their walls and forget their inhibitions, allowing for a conversation they would not typically have.
• The Type of Activity It Would FacilitateTo take advantage of the environment Bedtime Visions is set in and facilitate a conversation that a couple would not typically have, I chose the focal activity to be couples having conversations about their future together.
• The Language It Would EmployIn order to ensure that Bedtime Visions could be prototyped, I used Dialogflow, Google's tool to "build natural and rich conversational experiences." With Dialogflow, I was able to design Bedtime Visions, while considering the natural language processing capabilities of Dialogflow.

To test Bedtime Visions, I created a video of two couples—one old, one young—interacting with Bedtime Visions. Both couples in the video participate in the same activity by creating a vision of their future together in ten years. While each scenario isn't exactly the same, a user can recognize similar desires both couples have.

Sample dialog from Bedtime Visions's concept video

The video screen is split so that the younger couple occupies half the screen and the older couple is shown in the segment. Audio was also split between the left and right channels so that the viewer hears one couple in one ear and the other couple in the other when wearing headphones. At the center of the video is a Google Home, the device Bedtime Visions lives within.

Video screen split in half (older couple on the left, younger couple on the right)

Video focuses on the conversation of one couple at a time (audio isolated to to the right or left channel based on the side of the screen that couple occupies)

Bedtime Visions user sentiment was more positive than that experienced by Intimately and Curb. While one participant questioned how it was different than regular pillow talk, most participants saw it as a way to "set future goals", "strengthen... [a couple's] connection," "have fun together," and "empower... [a couple and their] future." One participant found it to be "weird" as they were "not used to robots asking these types of questions." Other insights from participants include:

• Bedtime Visions could support couples when confronting difficult topics.When Bedtime Visions encounters a difficult topic for two partners, Bedtime Visions could draw from a variety of frameworks that deal with such topics, thus helping the couple to work through this discussion.
• Bedtime Visions could be a part of a larger platform.Bedtime Visions could be a part of a large ecosystem of conversations that supports a couple in a variety of settings to effectively integrate into the everyday life of a couple. For instance, Bedtime Visions could leverage the specific conversations couples have in a kitchen or during a car ride.
• Bedtime Visions could encourage conversations beyond visioning exercises.Like the other two concepts, Bedtime Visions could confront numerous challenges affecting behavior change if it was limited to just visioning activities. If it also involved reflective or appreciation activities, Bedtime Visions would have a higher chance of affecting behavior change.

Unlike Curb and Intimately, Bedtime Visions helped me create a scenario centered on an artificial agent that was less familiar to users and take a deeper look at the possibilities for a set of humans and an artificial agent to achieve conversational symbiosis. What follows is a set of principles that were derived after testing Bedtime Visions.

• Bedtime Visions should facilitate impromptu and dynamic conversations.Bedtime Visions should not be limited to pre-planned conversations about a couple's future. It should also provide the opportunity for couples to have conversations that are unique to a place and time, for it to alter the practices of couples beyond a single interaction.
• Bedtime Visions should facilitate specific conversations for particular stages of an intimate relationship.A conversation set ten years in the future may or may not be appropriate for a couple who has been dating for a couple of months. Bedtime Visions should consider the context of a relationship and allow for conversations that fit that context, to effectively share frameworks and concepts with couples.
• Bedtime Visions should support and not direct a couple.Bedtime Visions should guide a couple through lines of conversations that it deems as beneficial for that couple. It should not force a couple to have a conversation about a vision of the future that the couple does not find interesting or worthwhile, so that a couple is able to evolve with the agent over time.
• Bedtime Visions should elicit imaginative and unfamiliar visions.Taking advantage of the environments in which Bedtime Visions is situated gives a designer the opportunity to enable conversations that a couple may not typically have within the context of their relationship, supporting greater self-disclosure between partners than what currently exists.
• Bedtime Visions should evolve its voice based on a couple and their communications.Like Curb, Bedtime Visions could evolve the language and delivery it employs based on the couple, and the context of their situation, so that it is able to productively communicate in a variety of situations a couple may encounter.

Storyboards depicting potentially unfavorable directions involving Bedtime Visions focused on the following developments.

• Bedtime Visions failing to create a neutral space.
• Bedtime Visions being biased towards certain visions of the a couple's future.

Through evaluative research, I recognized the opportunity for Bedtime Visions to support a larger system of conversations that supports a couple in a variety of settings, especially those that are ideal for certain conversations. For instance, one could build an iteration of Bedtime Visions specifically for a bathroom that supports the verbalization of those inner conversations someone has with themselves while sitting on the toilet, or an iteration designed for the kitchen as a couple decides what and where to eat.

An iteration of Bedtime Visions that supports a couple in a variety of settings