Rolling Chairs, Wagging Cars, & Designing Ritual Interactions for Autonomous Cars
Human Machine Rituals for Robots & Artificially Intelligent Systems
Written by Kursat Ozenc and Margaret Hagan, Contributor David Sirkin
This past January, we ran our Stanford pop-up class called Human to Autonomous Car Rituals at Stanford’s VAIL — a leading automative research lab. We are working on a larger academic study coming out of the workshop, but here we have a quick recap of what we are learning about how people might relate to cars in the near future.
Our class began with a central question: how will people build and conduct relationships with autonomous cars? We have a new challenge in human-machine relationships. As machines (like cars) become robots with the emergence of artificial intelligence, we don’t necessarily know how to interact with them. We lack norms and interaction patterns to develop meaningful relationships. Autonomous cars are an excellent example of this challenge. We are intrigued by them, yet not sure how to affiliate with and trust them.
We worked with students and professional designers to define what ritual interactions can help people navigate this new relationship with their car. After the sessions, we evaluated the proposed rituals to identify which are the most meaningful, and how they might be incorporated into future autonomous vehicle design.
We designed the workshop as a two-session sprint. The first session was an introduction to the autonomous cars and ritual design landscape with short lectures and design exercises. Second day was the design and prototyping day. Students sketched and prototyped rituals for their select situation.
We scaffolded the autonomous car rituals topic by gradually increasing the complexity of design challenges. First, we gave them a 5 minute wellbeing ritual challenge where they sketched car rituals similar to crazy eight (Google Design) method. We then gave a case challenge for a driver who is suffering from a long road trip. These exercises led to the main challenge of designing rituals for various relationship moments.
What were these relationship moments? We see these moments as a trust building journey. A trust building journey has several key moments: 1.Getting to know your car; 2. Routine building(commute); 3. when car makes a mistake; 4. When driver has a bad day; 5. When someone outside the car or another car makes a mistake; 6. When you ride with a stranger; 7. When you depart from your car. These moments are critical as they can build or break the trust between a driver and her car.
Students formed teams based on these relationship moments, and created rituals in three design cycles. In each design cycle, we ask students to try out a different relationship dynamic with their autonomous car. These are car as your dog, car as your boss, and car as your best friend forever. While giving these prompts we provided them the ritual design canvas and used user enactment as a way to prototype their rituals.
Students prototyped 7 rituals on how a driver or rider could interact with an autonomous car. Based on these rituals and the process, we discovered 7 insights on the metaphors, interaction modalities, and prototyping.
01 Need for More Animalistic Metaphors to Explore Interactions
Gesture was an interesting modality, when it came to cars communicating to people. Especially with the pet metaphor, of the dog, several of the groups created new interactions around the car communicating and building relationships with the human through the car moving parts of its body. For example, the windshield wipers could switch back and forth like eyelashes blinking, to show that it wanted to engage. Or at an intersection, the car’s antenna tail could wag to show that it was just waiting during the other person to go forward, almost like a wave. Or the cars ears, in the form of its side view mirrors, could rotate back and forth to welcome a person who was approaching the car. Or even more, the car could bow down, almost like a kneeling bus, to welcome its new Rider to get in, and to show that it was deferring and trustworthy, by making it easier for the person to get in.
In this respect, we found the pet metaphor to be very inspiring, to think about these relatively lightweight and quick interactions to build a relationship with the human. In fact, we think it would be worth in future workshops to imagine different types of animals in her actions. Not just a dog, which is the pet that we had explicitly named in our prompt, but perhaps a horse, a cat, dinosaur, or another range of animals to Think Through new interactions as well as new parts of the car that could be used to communicate with people inside and outside the car.
But if you make it physically act like a dog, will people sit on it? Will they they be wary of some types of functions we actually want them to take — because they make it into an animal, and apply limits/emotions/cognition to it?
02 Start with Pet Metaphor, grow it to a friendship
In terms of Personality, many of the participants thought that the car would best to begin as a pet like character, and then move towards becoming more like the best friend forever or a co-pilot. What they could see from their interactions, was that they wanted the first experiences with the car to be not that intimidating, and to make them feel like they were in control. The getting-to-know-you rituals were mostly about showing kindness, loyalty, inability to trust. That’s where the pet metaphor really came in useful. With the car acting like a dog, using physical Motions like wagging its tail, switching its windshield wipers, or making small sounds to reassure the person that it was trustworthy and that it was serving the person. But after relationship developed, the participants expected to build their trust with the car to the point that they wouldn’t want to be in so much control, but rather to defer the control to the car. They want the car to still care about the human, and predicted needs, and Decatur to its wants, but overall they would want to see more of the decision-making to the car, as if it were an equal or also perhaps a boss.
03 Mind the Anti-Experiences
When students tested some of their rituals, there was a pushback from the testers. Some of these rituals go beyond what’s acceptable or expected. In a way designers push the norms, to stretch or break them. We call these boundary situations.
In our enactments, especially around the dog scenario, we found that several groups ended up in interactions that were very uncomfortable, if not horrible for the user. For example, one of the groups created a sniffing ritual, where car decides who is the friendly rider, and rejects the ‘unfriendly’ one. This was uncomfortable and humiliating for that user. It also definitely brings very touchy subjects such as implicit and explicit biases that AI might have to make decisions. In one of the boss scenarios, the car acted like a boss to its Riders, who would be rideshare riders and not owners of the car. If the rider did something wrong inside the car, like spilling a drink, the car would eject it from the car. Again this was an anti user experience, in which they felt completely disempowered if not assaulted. In another scenario, when the car was playing the boss, the car would not listen to the rider, to slow down or just stop. This made the testers very uncomfortable, and worried for their safety.
What do these boundary scenarios mean to the actual behavior design? User enactments are a great method to uncover the boundaries of human machine relationships. They help designers to unveil the boundaries that they can work with, and make them aware of values and preferences of their users.
04 Isolating bubbles, and need for empathy, and social/civic connection
Many of the first wave of ideas are all around making the individual person super comfortable. These ideas trended to making the car into the ultimate Personal Assistant, Butler, or Servant — it was a one-way relationship, with the car expected to anticipate all the person’s needs, serve them fully, and offer even more services to solve the person’s problems, take care of all their tasks, and make them feel good.
There was definitely a trend towards narcissism and serving the human. That seemed to be the most natural tendency, when it came to interactions to create, but something like a mobile phone that was around space and surround experience. But upon reflection, many people expressed concerns that this would lead to more isolationism. This makes us think that in prototyping new interactions, we should expect the first wave to be all about complimenting if not cuddling the human, and we should challenge ourselves to think about longer-term consequences about how this might affect the person’s narcissism, or isolationism in society. Do we really want to be adding one more bubble to make a person insulated from the rest of the world?
05 Interest in Car Love Affair/Best Friend Forever made explicit.
We were surprised to see how many of the rituals were about falling in love with the car. We knew from our inventories of past relationships, that people often have very close relationships with their cars, but we didn’t necessarily think that people explicitly would want to fall in love with Car, or the personality the car was taking on. Some of the rituals very quickly turned into a situation like the movie. Where the car could expect at your every need, know how to interact with you, and be one of the best experiences of your day. To the point where some of the participants were imagining literally falling in love with this car, and the cars personality than traveling with you everywhere.
06 Moving Chairs and Lengthy Scene for Prototyping
Having rolling chairs and a long length of Space is really good for prototyping. We had gone in thinking that maybe we could have stations, but then we realized that having motion and rolling things with one of the keys to body storming new interactions.
07 Materials to Support more Car Gestures + Communication.
In future workshops we would provide more props and staging tools to allow for the car to communicate in non-verbal ways. This includes windshield wipers, lights, something like a tail or an antenna, or other mechanisms in which the group’s could prototype car gestures better. This was a limit in our team’s work, or some teams wanted to create it, but instead just had to verbally say the motion, instead of performing it. We would also like to encourage this nonverbal communication of the car, because the easiest thing to prototype is a chatbot like conversation, but several groups of potential in more gestural or hectic or motion ways of communicating, that got the point across without too many words or back-and-forth conversation.
We thank all our Human to Car Rituals class participants,VAIL staff and PhD students to host us during our popup class. Shout out to Professor Pablo Paredes for sharing his research on wellbeing in autonomous cars.
