Who if your future driver?
Speculative Design Service Design Interaction Design
What if we remove the human being from
the driver’s seat?
It is predicted that autonomous cars will be on our roads in the next five years, sharing economy for transport will become the norm and our most our daily human interactions will be replaced by AI.
Our design brief, at Royal College of Art, was about designing a service and its future interactions in the mobility space. But my team and I decided to take a speculative and provocative direction on the topic: Our research found that companies, engineers, and designers are primarily focused on building the technical requirements necessary to create functioning driverless cars, such as navigation and artificial intelligence (AI) systems, peer-to-peer vehicle communications, and advanced LIDAR sensors. But…
What are the implications of removing the human driver? What is missing from the equation?
To explore our questions, we designed our future AI driver, a voice interaction service that will replace the human driver in the car. We imagined multiple scenarios centered around an AI driver and filmed them to instigate a societal discussion.
By leveraging the power of the future users to shape the demand and need for autonomous vehicles, we aim to open a dialogue: How do we design the right AI driver service for autonomous cars?
The outcome of the project was a series of short films pointing out different interaction scenarios that were published online and exhibited at the Work in Progress Show at Royal College of Art. We asked the visitors of the website and the exhibition to write down their thoughts, emotions but also their memories of the driving experience.
The research we conducted was centered around the identified current market dynamics and captured the vision of the leading industry players. It is clear that traditional auto-mobility companies are being disrupted and we are talking for mobility as a service. Analyzing current signals and the social, economic, technological and political context we defined the trends and mega trends shaping the industry in the next 5-10 years.
It is evident that autonomous technology and shifts in economic models, are transforming the entire market. Companies are forming partnerships through acquisitions and new more intimate vendor relationships while investing enormous funds for R&D of next-gen mobility services and infrastructure. While each leading stakeholders has a different vision, from Tesla's Hyperloop to Uber's flying cars and Mercedes integrated mobility services, some common principles and trends arise providing the landscape for innovation.
Drivers of Change
The Role of the Driver
We conducted in depth interviews with families commuting in London, Uber drivers and employees who worked at bike/car sharing businesses to get a better understanding of current mobility problems we could solve for the future. The main insights we gained were:
1. A driver plays an important role in the car. They sometimes feel like they are therapists for their passengers and they also clean the car and maintain it after certain rides with messy passengers. This made us wonder what our interactions would be like with a car when riding alone and if cars were shared who would have ownership and take care of it?
2. Sharing a driverless car with strangers will create different social interactions. When travelling the driver is often the one with all the decision making power. Without a human driver who who will control the music/temperature? Will we each have our own AI assistants for the car and will we pay a premium for our preferences?
3. Safety is a huge concern. When we talked to an 8 year old boy Yael, he said he would be too afraid to go into a self driving car and couldn't imagine travelling without his parents. His parents and other parents with young kids felt the same. This opened a up a lot of concerns about the safety of autonomous cars especially for kids. Do we need to set up parental controls for the car? What happens if a car is hacked and how will an AI driver react when something goes wrong?
We realized that travel is messy and things don't always go according to plan. Although driverless cars will reduce accident rates and congestion on the road, how will people react on the few occasions something does go wrong?
At the end of our research we had more questions than answers. How could we design the best AI driver when there were so many social situations we needed to address? Instead of designing a solution, we decided to create a series of videos to start visualizing what these situations could look like to gain awareness and start a conversation among our network of designers.
We selected a few scenarios we wanted to highlight, shot videos and exhibited our project at the Work in Progress show at the Royal College or Art. Based on the interactions we highlighted here are some of our recommendations and our first stab at answering: How do we design the right AI driver service for autonomous cars?
Scenario 1 | Riding Alone
Here we explored what it would be like to interact 1:1 with a driverless car. We came to the conclusion that the best experience when riding alone is a personalized AI assistant that can speak to you and also answer questions you have about the journey. Experience is very important - the car has to get your destination right and be able to listen to any requests you have along the way and adjust its route accordingly. We can’t yet tell how much a passenger will actually talk to the AI in the car but we will probably start to see some interesting behavior once autonomous cars are in use.
Scenario 2 | Sharing a car with strangers
As the sharing economy is on the rise, we predicted that there will probably be a shared autonomous car services. Unlike the train or the bus, a car has a fewer number of people in an intimate space. We think without a human driver this could create some tension between passengers if they want their preferences in the car. We recommended in this situation there needs to be some hierarchy of power. Maybe the first person in the car gets their choice of music and temperature or people can pay a premium for their preferences for the ride. Additionally the AI could be a type of mediator that enforces default settings if there is a conflict and report any issues in the car to an authority if something nonconsensual happens between passengers.
Scenario 3 | Getting into an accident
Although it will be rare, there will be a few occasions when an autonomous car will get into an accident. Without the human by the wheel this will be scarier because the AI will be making decisions on your behalf for your safety. One solution we had in this instance was linking the car to a support line that had an actual human that could communicate with you and guide you what needs to happen to stay safe in an accident.
Scenario 4 | Driverless cars and young kids
Here we explored how an AI voice assistant might interact with a kid. We realized that there needs to be some sort of parental controls for very young kids so that they don’t mistakenly ride alone in an autonomous car without their parents. Parents felt very strongly about this and wanted to make sure that they were always with their younger kids and would not trust a driverless car. Additionally as with other online content parents expressed that they would like to make sure their kids had a good balance of educational content and also being offline from the internet when travelling.
Scenario 5 | AI as your personal assistant for all travel
Here we explored an AI that navigated you when you walked, biked, drove or took public transport. It knows all your preferences and in fact knows you better than you know yourself. Would you be comfortable with it making decisions for you? In our video we show an example of your assistant taking you through the scenic route so that you can get more exercise. When we shared this video with others we realized that when we design such assistants we have to create a very fine balance between a AI that asks your permission vs an AI that makes decisions for you.
Through our research and outcomes we realized that with every transformative technology there are side effects. So we decided that designing a solution for a perfect AI driver service would not be a realistic goal for our project. Instead we used design to create a conversation to get people to think about context and social interactions when it came to autonomous cars and not just a technological solutions. At the the exhibition we collected more feedback/stories from visitors that were both positive and negative and it really helped us broaden our views of positive/negative social interactions that could happen in a car and help us come up with better recommendations for an AI driver service.
Learnings & Reflection
The traditional design process for products and services as well as the definition of service design is being challenged in this process. It was evident from the very start of the project, that in order to design a service for the future we need to transport ourselves in this context and build our future world before building the service itself.
Our design process began with a wide analysis of the drivers that shape the future using the PESTLE framework, while at the same time we strived to understand the current market forces that shape the industry. Our socioeconomic and trends analysis were the foundation of our world-building exercise. We developed future scenarios that lay in the scope of possible alternatives1 and within those scenarios illustrated the future world.
It was in this phase when we came across the implications of AI in our lives and mobility leading as to our core question; what happens when we eliminate the human driver from the car. The above question, informed the second round of research, aiming to talk and understand real people’s feeling and needs, listening to their personal in-vehicle stories and comprehend the demands of extreme users such as children, people with disabilities and older people. We conducted in-depth interviews and observed the behaviour of people and families in the streets and residential areas.
Our findings demanded that we follow Dunne and Raby’s approach to Speculative Design, aiming to provoke and open a discussion through our project rather than design a feasible service. The short films created were exhibited at the WiP Show at RCA where the visitors were able to interact and discuss the issue. The final project, in the form of provocations, as well as the report generated, was delivered to Moovel Lab, a Daimler Group Company.