The funds are open to both full-size autonomous vehicles and city-centre low-speed vehicles.

Guest Post: Iris Behind The Wheel. How Will We Feel When Cars do the Driving?

Here at Transport Evolved, our editorial remit is Cleaner, Greener, Safer, Smarter, which means that we cover everything from plug-in cars and hydrogen fuel cell vehicles through to advanced safety features and safe driving cars. 

Last year, Nissan LEAF owner and life-long Psychology academic Stephanie Lay wrote us a fascinating four-part series on the psychology of electric car range anxiety. We — and you — loved it so much we’ve asked her to do the same this year on a different subject: autonomous vehicles. 

With more self-driving car studies and test programs now being set up than ever before, Stephanie examines why some of us feel a little uneasy about the thought of being driven by a computer program rather than a human. 

“Hello, I’m Iris. I’m your car’s automated system… you can remove your hands from the wheel and your foot from the pedal because I’ll be driving“ From Waytz et al (2014).

What's behind our reaction to autonomous vehicles?

What’s behind our reaction to autonomous vehicles?

Whether they’re described as self-driving, driverless or robot cars, autonomous vehicles are getting a lot of attention in the popular press as they start to venture off their test tracks and on to roads and walkways with the rest of us. A series of trials are underway in the UK, including where I live here in Milton Keynes. I’m a psychologist carrying out research into things that are almost but not quite human and I also have a keen interest in new vehicle technologies so it’s very exciting that one of the trials is on my doorstep. Later this year I’ll be looking out for the LUTZ Pathfinder pods as they start to make their way around the town, albeit still under carefully supervised conditions. It’s part of an exercise to evaluate how feasible they may be in the real world and raises some very interesting questions about the nature of transport and how we feel about the vehicles we inhabit and encounter.

Autonomous drive vehicles have been hailed as the future of motoring by taking the error-prone humans out of the driving seat and turning long commutes into time for work, relaxation or entertainment, all the time allowing more efficient use to be made of the limited space available on our overcrowded roads. For that to happen, this new technology will need to be accepted and normalised and a lot of drivers are going to have to get used to being passengers. What happens to our feelings towards cars – not just ours, but everyone else’s as well – when we’re no longer in the driving seat?

There are many psychological factors that could cause autonomous vehicles to be rejected, concerning everything from generalised anxiety over the risks of new technology, a loss of the parts of your identity that are caught up in the concept of yourself as a driver or an unwillingness to cede responsibility to something that makes decisions for itself, through to fears that the legal system is unprepared for the arguments that might occur with driver versus driverless vehicles in collision. This is a relatively new area for psychological research so in this article, I’d like to focus on two studies published in the last year, which look at two factors influencing the acceptability of autonomous vehicles.

Self-driving cars: what's your opinion of them?

Self-driving cars: what’s your opinion of them?

The first study starts to look at some of the compromises that are going to become necessary if autonomous vehicles become commonplace. The research was led by Scott Le Vine at Imperial College London, who identified that there will always be a tension between the costs and benefits for the individual traveller and for the efficiency of the transport system as a whole. If they are to succeed, autonomous vehicles should allow detailed control of the vehicle’s behaviour by adjusting aspects of driving performance to meet different requirements.

This study compares the two specific variables of comfort and speed, as travellers want to enjoy the most comfortable journey possible while getting to their destination as quickly as they can. Passenger comfort can be affected by how quickly the vehicle accelerates and decelerates and Le Vine et al used examples from passenger trains where high-speed trains tend to use smooth and gentle transitions and local light rail tends to use jerkier, harsher braking and acceleration. Journey speed is dependent on, among other factors, making efficient use of road junctions to get as many vehicles through as possible. Le Vine and his colleagues used computer simulations to model a road system, including a junction, where autonomous vehicles were sharing the road with human-driven vehicles.

They varied the different acceleration and deceleration profiles to create carefully controlled scenarios representing different levels of passenger comfort, and measured the effect each one had on the capacity of vehicles passing through the junction. They found that when the autonomous vehicles used the smoother, more comfortable driving profiles this actually had a negative impact on the efficiency of the traffic flow compared to the less-passenger-friendly jerkier profile as fewer vehicles were able to pass through the junction in the same amount of time. So, counter-intuitively, it might be necessary to make choices that reduce the comfort of the ride experience in order to get the best advantages in traffic flow efficiency.

Test schemes like this are integral to understanding the psychology of accepting  autonomous vehicles on our roads.

Test schemes like this are integral to understanding the psychology of accepting autonomous vehicles on our roads.


This would seem to go against the idea that autonomous vehicles will inevitably give us the ability to relax or work while the car takes care of the driving and presents a more complex balancing act between different factors. Le Vine et al acknowledge that this is an initial study designed to stimulate further research and discussion, and the real world experience of being a passenger in an autonomous vehicle would be based on much more than those two factors. They conclude that further research is going to be needed to understand these new ways of thinking about and relating to our vehicles if autonomous vehicles are going to be generally accepted.

Le Vine’s study was grounded in the mechanics of how autonomous vehicles can be programmed and how the behaviour of individual vehicles can affect overall traffic flow. The second study, led by Adam Waytz at Northwestern University in the US, looked at autonomous vehicles themselves and asked what might help people to trust them to take over the task of driving, a skilled and complex learned behaviour which is performed at speed under potentially risky circumstances. They noted that people will put more trust in a behaviour that they believe to be carried out intentionally and mindfully than one that is haphazard or thoughtless.

An artificial agent that appears to possess a mind is perceived as more trustworthy, competent and safe than one acting according to preset rules, even if the actual behaviours themselves are identical. Anthropomorphism is used to describe the attribution of humanlike qualities to something non-human so this study set out to test whether trust in autonomous vehicle would increase when it was given humanlike traits.

Autonomous  vehicles: do you welcome them or fear them?

Autonomous vehicles: do you welcome them or fear them?


They tested this in a driving simulator where people were asked to drive an autonomous vehicle that could control its own steering and speed when driving on the open road but required human assistance for low speed manoeuvres. (Simulations of driving a non-autonomous vehicle were also included for comparison.) Some of the participants were given anthropomorphic cues about the vehicle by giving it a gender and an identity. It was also given a ‘voice’ where information about the controls and the journey was given as if spoken by the car using pre-recored voice files. The driving course was identical for all participants so the only aspect of the experience that varied was the human-likeness of the car. The results were striking as participants who drove “Iris” liked ‘her’ significantly more than participants who drove the same car without the cues. They also trusted the car more, which was measured both by their own reports in a questionnaire and also in a novel way where an unavoidable accident took place during the simulated drive.

The participants in the anthropomorphic condition were not as startled by the crash and they were less likely to blame their vehicle compared to the other two conditions. Waytz explained this as meaning that participants had attributed thoughtfulness to the anthropomorphic car because of its humanlike qualities and therefore were less likely to think it could be to blame for the accident. Again, all participants had had the same experience during the simulation so these differences could only be due to increasing the humanlike nature of the car by giving it a name, identity and a voice. They concluded that when we anthropomorphise something, we’re doing more than just believing that it seems human, we believe it is capable of sharing our human traits of being able to think and feel. To make a technology more trustworthy, it must seem capable of possessing a mind.

They may feel alive, but autonomous cars are just very clever computers.

They may feel alive, but autonomous cars are just very clever computers.

The study suggests that to make people more willing to engage with autonomous vehicles it will be necessary to make them more human. However, there will be a limit to this and my own research into the uncanny valley has found that entities that start to gain those humanlike properties but can’t give a fully realised and realistic impression of being human are quickly rejected for being too eerie. Iris only presented rudimentary human cues so it didn’t stray too close into unsettling territory. The eeriness barrier is going to be a hard one to overcome in the development of autonomous vehicles, for example, it might be possible to accept that machines can make better decisions for us on matters of logic but it could seem very unsettling to feel that not only is your autonomous car a measurably safer driver than you are, but that it also gives the impression of thinking about it as it does so.

Happily, we have time to work on these issues. The current pilots in the UK are starting small, with trials in just three cities for the time being. These are just the first steps in long-term development projects. Steve Yianna is in charge of Transport Systems Catapault who build the LUTZ pods that will soon be exploring the streets near to where I live, and he believes that we’re at least 10 years away from the point where we can let the autonomous vehicle take over and drive while we sit in the back to read a newspaper. This is the perfect time for psychologists and car manufacturers to learn more about what might influence the take-up of autonomous vehicles – research such as the studies above will greatly enhance our understanding of the human response to this new technology in preparation from when they’re eventually out and about on their own.


Le Vine, S., Zolfaghari, A., & Polak, J. (2015). Autonomous cars: The tension between occupant experience and intersection capacity. TRANSPORTATION RESEARCH PART C, 52(C), 1–14. Elsevier Ltd.

Waytz, A., Heafner, J., & Epley, N. (2014). The mind in the machine: Anthropomorphism increases trust in an autonomous vehicle. Journal of Experimental Social Psychology; Journal of Experimental Social Psychology, 52(C), 113–117. Elsevier Inc.

About the ‘UK Autodrive’ autonomous vehicle trials.


Want to keep up with the latest news in evolving transport? Don’t forget to follow Transport Evolved on Twitter, like us on Facebook and G+, and subscribe to our YouTube channel.


Want to keep up with the latest news in evolving transport? Don’t forget to follow Transport Evolved on Twitter, like us on Facebook and G+, and subscribe to our YouTube channel.

You can also support us directly as a monthly supporting member by visiting

Share on FacebookTweet about this on TwitterShare on Google+Share on LinkedInDigg thisShare on RedditEmail this to someonePin on Pinterest

Related News

  • Jeffery Lay

    Many people anthropomorphise cars already, giving them names and pseudo-identities. I wonder how this affects their perception of the vehicle’s part in their driving, even more the involvement of autonomous drive systems?

  • Chris O

    Great example of an autonomous car with anthropomorphic cues: the hilarious Johnny Cab in the Sci-fi movie Total Recall :nnn Schwarzenegger quickly realized that autonomous driving was not going to work for him:nnn think in complex (urban) environments people will be quickly inclined to turn off a slow, overly cautiously maneuvering AI and do the job themselves.

  • Joe

    I recently bought a robotic vacuum cleaner and have found myself treating it like a person, or more accurately, as a pet. Beyond being silly or bizarre, this is a bit of a problem because when the device malfunctions, and it very frequently does, I am annoyed as though it has misbehaved rather than just being a poorly designed product. (Avoid bObSweep brand vacuums no matter how good the price is.)

Content Copyright (c) 2016 Transport Evolved LLC