The diffusion of the autonomous car

This is the introductory part of a series of posts where I want to explore how autonomous cars could diffuse. By diffusion I refer to the 

„process in which an innovation is communicated through certain channels over time among the members of a social system. It is a special type of communication, in that the messages are concerned with new ideas.“ — based on Everett M. Roger’s Diffusion of Innovations.

Further, I will refer to the autonomous car not as an isolated product, but as a product of a socio-technical system.

The first case (autonomous cars as an isolated product) would be similar to Ariel’s 3in1 PODS Washing Tablets. These are capsules which package all the “ingredients” you need for washing your clothes in predosed quantities.

The Ariel’s 3in1 PODS Washing Tablets

Although they are awesome (I 😍 them ) they will not significantly change our lives, call for new changes in our daily routines or envoke protests. They will make “only” make washing easier and quicker.

The same, however, is not true about the iPhone. The iPhone brought with it new business models, changed how run our lives and even caused a drop in chewing gum sales. It caused for a transformation of the socio-technical system of communication (computing/internet/phones might be other suitable categories).

Such socio-technical systems are used to fulfil societal functions such as transportation or energy supply and revolve around one core technology (e. g. cars) but include everybody (groups such as policy makers, consumers, manufactures) and everything (e.g. competing technologies, the overall market or user practice and symbolic meanings) involved in it [1].

In analogy to that we can either see the autonomous car as a product or one part in the overall socio-technical transportation system. If we see autonomous cars as a product we would just need the technology for autonomous driving which apparently is ready [2] and ensure that consumer adoption. However, judging by the autonomous car’s long-term potential [3] and what the introduction of cars in the first placed meant to society [4] autonomous car should be seen as part of the overall socio-technical transportation system. This system of which autonomous cars are part of can be best described by being a personalized, on-demand, door-to-door, automated transportation system with little not no accidents.

This being said, the autonomous car could then be seen as a means to an end. This in turn allows criticism to my approach. I am seeing the autonomous car as the central technology of the future of mobility. Considering everything going on today (e. g. sharing economy, increase of online shopping…) “the future of mobility”, instead of the autonomous car, could be the “right” unit of analysis. In this “future of mobility” autonomous cars would be just one part of the puzzle alongside bike sharing, flying cars, automated underground tunnels (e. g. The Boring Company’s electric skates) and drones.

As I am not doing this, lets continue looking at the transportation system where autonomous cars will be the central technology. In this transportation system we will eventually have driverless cars as opposed to self-driving cars. Whereas the car as a self-driving vehicle with a present driver will make our lives safer and more comfortable it will not overhaul our lives completely. On the contrary, driverless vehicles will change our lives completely [3]

Although, this transformation might make the implicit assumption that autonomous cars will at one point completely replace manual cars this mustn’t necessarily be true. In these series of posts I will elaborate on this.

Seeing the autonomous car as a solution then shows that diffusion of the autonomous cars will take “long”. To explain how “long” it will take one might use the Bass diffusion model. And I will do so at a later time. For now, I want to explain why it will take “long” and how the diffusion would be most effective.

The diffusion of autonomous cars depends on consumers’ perceptions, the influence of the involved technologies the influence of the wider landscape

To explain what will influence the diffusion of autonomous cars and why it will take “long” I use three categories:

  1. The perception of autonomous cars by consumers
  2. The influence of the involved technologies
  3. The influence of the wider landscape

The perception of autonomous cars by consumers

  • Consumer adoption of autonomous cars
  • The changing perception of autonomous cars in movies
  • How flexibility and comfort might influence the autonomous car’s adoption

The influence of the involved technologies

In accordance with the socio-technical view, autonomous cars must not be seen as exclusively dependent on self-driving technology. A plethora of other technologies is going to influence its diffusion.

How and which technology might influence the autonomous cars’ adoption

The influence of the wider landscape

  • Ecosystem
  • How companies’ strategic actions might influence the autonomous cars’s adoption
  • How legal and privacy might influence the autonomous cars’s adoption
  • Transport & geography — literal landscape changes

The diffusion of autonomous cars will start in niches

Companies have two options for rolling out their autonomous cars; they can either roll out each level of autonomy incrementally (like Tesla does) or in a bing bang approach start selling their cars only when they are capable of fully autonomous Level 5 driving (like Waymo does)

Whereas the bing bang approach lowers the risk of accidents significantly, incremental rollout would — although it might sound counterintuitive at first — speed up the diffusion. This is because incremental rollouts give consumers

  • earlier access to self-driving technology
  • earlier access to less radically different self-driving technology than they are used to
  • earlier access to less radically complex self-driving technology than fully autonomous self-driving technology represents

In other words, in an incremental introduction scenario, self-driving cars would look like a natural development. In regards to the above-mentioned adoption drivers this meant that consumers can try (trialability) and observe (observability) technology which is not too complex (complexity) but very compatible with what they are used (compatibility) [6].

Furthermore, incremental introduction allows for the ecosystem’s development such as adjustments on the job market as fully autonomous cars will erase some jobs but also create new ones [7].

Possible media backlash and their consequences (negative perception of autonomous cars) caused by imperfect autonomous cars speak against the incremental rollouts. Reports like this one by Consumer Reports would then be just the beginning.

Another disadvantage of incremental rollout is possible loss of driving practice. Assuming that people will let drive more than they will drive themselves, they will be out of their driving routine once they actually have or want to drive.

 → link 2 var tech + dynamics + intro w/ qute from bmw + listings

Several companies have announced to have self-driving cars by 2021:

Whereas we really might have Level 4 or Level 5 autonomy ready by 2021, one must not forget all the other technologies surrounding autonomous driving technology. I have looked at them here. →

Furthermore, I doubt that such a Big Bang approach where we immediately have Level 5 autonomous cars is realistic. I believe that the introduction will be an incremental one. The dynamics of this incremental introduction are visible alongside use cases and the technology itself. Here I have explored the various use cases and here I am trying to predict the autonomous car’s technical diffusion. Although I am dealing with them separtely, the are related because the range of possible use cases depends on the technical possibilites. 

Incremental introduction and Rogers

Incremental introduction and observability

In an incremental introduction of autonomous cars is lower than in a big bang immediate Level 5 autonomy situation. The disadvantages of such an incremental introduction are visible today. For example, a study published in Taking a Drive, Hitching a Ride: Autonomous Driving and Car Usage found that 44 % of participants hat had no knowledge of autonomous driving in 2016. Although this is just one study its reasonable methodology [16] allows for some generalization, at least for Europe. Although a lot of media perception happened since 2016 I doubt that number of knowledgable people has increased. This means that still approximately 40% of the population doesn’t know anything about autonomous driving, a technology that is supposed to hit the roads by 2020 [17]! In this sense, driving assistants (such as braking assitens) will be of utmost importance for demonstrating autonomous driving functionality and thus providing direct observability [3].

Will we ever have a fully Level 5 autonous world

Complete still stand without fallback in case of failure in a fully Level 5 autonous world on reason against ist

Complete stillstand without failback in case of failure in a fully Level 5 autonous world

In regards to autonomous driving software, i. e. the software that makes the autonomous driving decisions, it seems that each company is working on the own algorithms. In regards to V2V and V2X communication at least some common standards will be necessary. And in regards to maps, several companies are jointly working on them.

The use of similar software in mapping and communication and a possible interchange of autonomous driving software between manufacturers poses a dangerous dependence. If all cars of the future rely to the same extend on the same code, all these cars will be affected simultaneously. Whereas we have such issues with traditional cars as well, I believe that car-interconnectivity will make matters more complicated with autonomous cars.


[1] see Technological Transitions And System Innovations: A Co-evolutionary And Socio-technical Analysis

[2] Several companies have announced to have self-driving cars by 2020/2021:

[3] examples of the autonomous car’s potential:

  • The separation between people and cars is blurring: Mercedes imagines the fully autonomous future to be a place where humans are cars are interacting in closer proximity than today
  • changing role of the car: in a future with fully autonomous cars we could hold meetings in them or use them for family time

[4] Technological Transitions And System Innovations: A Co-evolutionary And Socio-technical Analysis does a great in explaining the changes cars brought with them, here are some of them:

  • the car enabled suburbanization by being able to driver wider and faster then with previous modes of transportation
  • it contributed to solving “overurbanization” (overcrowding and the like of cities) by enabling suburbanization
  • it lead to the rise of recreational travel
  • it lead to the rise of individual, on-demand transportation
  • it lead to the rise of a car culture (e. g. shopping malls at edges of cities, fast food restaurants at highways and drive-in movies)
  • the street’s role changed from social gathering places to transportation arteries: whereas before the car people were socializing (e. g. children playing) on streets, streets are now mostly car domain and people are pushed to dedicated areas (e. g. playgrounds for children)

[5] MLP is based on Technological Transitions And System Innovations: A Co-evolutionary And Socio-technical Analysis

[6] I have explained these influencers of innovation adoption here in detail:

[7] I have explained the that ecosystem’s parts here: