Table of Contents

There are three groups of questions that are critical to preparing for the availability of CAVs:
  • How quickly would they be available for purchase and use? This is likely to depend on factors like the speed with which technology can progress from Level 2 to 5, the premium to be paid for each level compared to a conventional car and to the type-approval and regulations imposed by authorities.
  • How people will acquire and use them, in particular, the choice of owning or hiring. This will probably depend on the relative costs of owning and hiring CAVs (including insurance and delay to service), the attitude towards ownership, regulations regarding parking and empty running and other factors.
  • What will be the balance of community costs and benefits compared to the individual costs and benefits and what policies and regulations should be in place at the time to manage this balance? Most of the social benefits accrue with Level 3 (accidents, some capacity improvements) but most of the individual benefits accrue with Level 5 (freedom to undertake other activities, reduction of parking problems, lower or no insurance premium, opportunity to off-set costs by hiring out when not in use).

This site investigates these key questions searching, summarising and updating the range of answers that are most likely to be useful and relevant. However, in many cases there is no objective and certain way to estimate the timing and importance of these impacts and therefore it is desirable to have expert opinions. As a first step, the site is using the participants in a Delphi poll undertaken late in 2016 as the basis for an International Panel of Experts.

The Delphi Method, or poll, is a structured communication technique developed as a systematic, interactive forecasting method which relies on a panel of experts. Delphi is based on the principle that forecasts from a structured group of individuals are more accurate than those from unstructured groups.

This particular Delphi poll involved 45 international experts, all with more than 10 years of experience in transport and traffic forecasting. They were no technology specialists; rather they were concerned with the likely impact of these new technologies and therefore ideally placed to answer these key questions. To account for geographical and cultural differences they were grouped into 5 regions, the USA & Canada, Western Europe, Australasia, Latin America and the Rest of the World (RoW), mostly Asia and Russia:

Pic 4.png
Pic 4.png

They were involved in a total of three rounds of questions on their views on some key impacts of the early years of Levels 4 and 5 CAVs (as they would be the most disruptive) and how they would be influenced by the context, region and policies. The survey was undertaken in late 2016; it is possible that by now some of the answers would change given the knowledge acquired in the interim period. What follows is a summary of the findings.

It is recognised that the answer to these key questions would depend not only on the country itself but also on the region or area (urban/rural) within each. The current set of key questions that this Delphi poll and site addressed are as follows:
  1. When CAVs will be available for purchase by ordinary citizens in your country/region?
  2. When CAVs will constitute 10%, 20% and X% of the car fleet in your country/region?
  3. What would be the premium (in Pounds/Euros/US$) to be paid on purchasing a CAV compared to a normal car?
  4. What proportion of the CAV fleet will be owned by individuals whilst the rest is available for hire by the minute or with an Uber style pricing?
  5. How expensive would be CAVs used as MaaS compared to current options (Taxis, Uber)
  6. What will be the impact on the capacity of freeways/motorways (grade-separated junctions) when a specific proportion of total traffic is CAVs?
  7. What will be the impact on the capacity of urban roads (with traffic lights and roundabouts) when a specific proportion of total traffic is CAVs?
  8. What will be the effect on trip making for CAV owners and for those using them as MaaS?
  9. What will be the impact on urban bus Public Transport demand when CAVs are X% of the car fleet?
  10. What will be the impact on Urban Fixed Track Public Transport (Rail, metro, LRT, BRT) demand when CAVs are X% of the car fleet?
  11. What will be the effect on the behavioural (for demand modelling) Subjective Value of Travel Time Savings (SVTTS) on the part of CAV users?
  12. What should be the change in the Social Value of Time (for use in Cost-Benefit Analysis) to be assigned to CAV traveller?
  13. What will be the impact on traffic accidents at different levels of CAV share of traffic?
  14. What will be the most important social or cultural drivers that will affect the choice of owning or hiring CAVs in the future?
  15. What technical aspects of CAVs will influence the decision to hire or own them and in what direction?
  16. What Economic or Pricing issues are likely to influence the decision to hire or own CAVs and how they are used?
  17. What transport policies are most likely to influence the decision to hire or own CAVs and how to use them?
  18. What legal aspects of using CAVs are most likely to influence the way they are used?
  19. Are there any ethical considerations to bear in mind when developing policies on the use of CAVs in the future?
  20. What are the environmental considerations to bear in mind when developing policies on the use of CAVs in the future?

The main objective of this Delphi was to gather views on the best way to approach transport and traffic forecasting for a future with Automated Vehicles (AV) level 4/5 (that is they can operate without a passenger/driver) in circulation. There was no assumption about interconnected vehicles or a particular level of intelligence in the infrastructure.

Overall, there was more dispersion than consensus of views reflecting our limited understanding of how the future will pan out. Just taking the average response will not be sound. The table overleaf shows a numerical summary of all responses including the standard deviation. One should not adopt at face value the mean response as an “average forecast” and it is critical to bear in mind these are only subjective estimates from this panel:


We now look at each question in turn in greater detail. The graphs and comments below do not distinguish different regions; they are all aggregate of the responses.

1. What year do you estimate AVs will be available for purchase by ordinary citizens in your country/region?

The mean for all regions is 2023 but the US and Canada expect them to be available by 2021, incidentally a year earlier than in Round 1. Europe continues to be more pessimistic (average 2025) and Latin America expects them even a year later; this looks plausible. The following chart shows the cumulative distribution of answers to that question.
Pic 5.png

This time it was possible to specify availability before 2020 (the lower value in Round 1) and two respondents used that opportunity arguing that recent events justified this.

According to the responses, it seems fair to accept that, for the purpose of travel forecasting, AVs would be available by 2026 at the latest. Effective availability, and more importantly the date when AVs will be 10% or more of the fleet, will vary from country to country.

2. What year do you estimate AVs will constitute 10% and 20% of the car fleet in your country/region?

This question is also related to the rate of growth in car ownership in a region and to the rate at which cars are scrapped either because of age, accident or export.

The mean response for the 10% threshold was 2032 but with a large dispersion (standard deviation 7 years). North America expects this to be achieved 7 years after AVs are available; the expected lag in Europe is 8 years but in Latin America is 13 years. The Rest of the World, heavily influenced by Asia, seem to expect to reach this threshold faster: 4 years.


As can be seen, those who expected early availability also expect AVs to reach 10% of the fleet earlier, even as soon as 2020 (just 2 years later). This reflects different views on how attractive AVs will be, and the role of mobility companies like Uber in adopting the technology and replacing drivers ‘en masse’.

This seems to be a fair comment as companies like Uber have significant influence on the purchase of new vehicles and can increase the AV share of the fleet very rapidly. On the other hand, Uber-like firms are not successful everywhere and are absent in many areas. This suggests that there will be differences in the rate of AV penetration not only among countries but also within them; it has been suggested that in world cities like New York, San Francisco, London, Paris and Hong Kong, AVs will reach these thresholds faster than in other locations. For 20% of the fleet, this seems to be expected 5-6 years after they reach 10% of the fleet, reflecting an acceleration of the trend to adopt them. Regions seem to be consistent in the time it takes to move from 10% to 20% of the fleet. The mean year for 20% AV penetration is 2037 and by then the expected impact is significant as shown later.

3. What would be the premium (in US$) to be paid on purchasing an AV compared to a normal car?

One would expect early AVs to require a premium that will be reduced as more vehicles are deployed. In this case, the focus is 5 years after they become available, close to when they constitute 10% of the car fleet, whether they are rented or owned. The question glosses over the problem that many AVs will not really be comparable to an existing vehicle as they may have facilities to work or relax.

There were many different responses, from next to nothing to $10K or even $15K. One respondent misinterpreted the question stating that he would pay nothing extra for this “feature”.

It is clear that there will be a premium and that this is not perceived as too different in different parts of the world. The overall average is US$6,700 (a bit lower than in round 1) but Australasia expects only US$ 5,600 and the US & Canada even less: $5,100.

Pic 10.png
This premium will be relevant for individual purchases but less so for mobility companies like Uber. The premium will be, by all accounts, less than a few month of a driver’s income in the developed world; the situation will be different in emerging countries where the cost advantage of AVs for mobility companies will be less marked (and the infrastructure and signalisation will be, perhaps, of a lower quality).

4. What proportion of the AV fleet will be owned by individuals whilst the rest is available for hire by the minute or with an Uber-style pricing?

It is recognised that this proportion may well change over time but this is one of the most critical questions affecting the impact of AVs on trip induction and public transport. During the second Round of this Delphi, Uber announced it intends to offer Volvo AVs to a sample of users this year in order to test the technology; they will be accompanied by an experienced driver in case it becomes necessary to revert to human control. It is likely that the Uber model will prevail and that this will make the rental of AVs much simpler and more attractive.

On average, respondents estimated that 42% of AVs would be owned. Latin Americans seem to be more attached to owning a vehicle and their average expectation is 56%. Western Europeans are in the other extreme, perhaps because they are used to better public transport, with only 33% of ownership estimated there.
Overall, more than 60% of the answers stated that the majority of the AV fleet will be available for rent. Indeed, Uber and Lyft are extremely well placed to exploit this technology with their experience to manage fleets and direct vehicles to areas of greater demand at any one time.

One possible interpretation of the figure below is that respondents are split between those that love their cars and driving, and those who care only about mobility as a service, be it public transport or Automated Vehicles; emotions influence strongly our views of the future.

Pic 12.png

5. For hired AVs, how do you expect the pricing to be set compared to Uber? Present it as a ratio over Uber pricing (ignore surges).

The ratio AV/Uber hire has a mean around 0.9 but this hides significant different views. Some respondents believe that the absence of driver will more than compensate the additional cost of purchasing and running AVs; moreover, AV insurance is likely to be cheaper as well. Others seem to think that and AV Uber will have to sort out large stabling and maintenance facilities and that maintenance costs, because of more demanding standards, will also be higher. After all, Uber/Lyft currently takes advantage of the ingenuity of their drivers to park and look after their vehicles.

We do not know what was behind those 13% of answers that expect AV hire costs to be higher than Uber.

6a and b. What do you expect will be the impact on the capacity of freeways/motorways (no at-grade junctions) when a specific proportion of total traffic is AVs?

This question is posed when AVs are 10% and 20% of the traffic; in other words, any induction is taken into account and therefore this threshold may be reached with AVs being less than 10% or 20% of the fleet. The higher the proportion hired rather than owned, the earlier this threshold will be reached, especially off-peak. For simplicity, we assume this 10% applies to the peak when capacity is a limiting factor.

Several respondents argued that at low traffic penetration for AVs the impact on capacity would be minimal or even negative, as they would only drive legally. North Americans are more optimistic about a beneficial effect on capacity.

The situation is different for a 20% AV presence in traffic. There the mean view is that at least a 10% improvement in capacity will be achieved at that level. There are, of course, extremes. Some believe that the impact on freeway capacity will be large; this is in line with the argument that even a small proportion of AVs will be able to smooth traffic, prevent shockwaves and therefore improve capacities significantly. Nevertheless, the most optimistic of the respondents (50% increase in capacity) was looking at the number of passengers moved per lane and argued that ride sharing, as supported by AVs, will increase this definition of capacity. As the question had not specified that capacity was meant to be in vehicles/hour/lane this is a valid interpretation.

7a and b. What do you expect will be the impact on the capacity of urban roads (with traffic lights and roundabouts) when a specific proportion of total traffic is AVs? Present this as a ratio over current per lane capacity.

I would have expected a different set of responses here on the grounds that traffic moves in a completely different way. However, it seems that the higher saturation flow expected from AVs would also improve capacities and that traffic lights would slowly adapt to the changing traffic composition.
On the other hand, it has been suggested that as AVs will drive and perform only legal movements they will not avoid problems in urban traffic as well as human drivers. Overall, the effect seems to be neutral at 10% and marginally positive at 20%.

8a and b. What do you expect to be the effect on trip making for AV owners and hirers when AVs five years after they become available? Present this as a ratio over current average vehicle kilometres travelled (VKT) per person per year.

We all would like to know whether traffic will increase or decrease when AVs replace cars. The global average indicates a 10% induction of new VKTs for AV owners and a 10% reduction in VKTs for AV hirers. This makes sense as one would expect some reduction in VKTs once the cost of hire is available up front compared with the marginal travel cost for a car owner that is seldom, if ever, known.

In contrast, the figures for the additional VKT empty vary so much among respondents that are really tricky to analyse. The average is 13% of additional kilometrage. The respondents are consistent in expecting that the extra VKT will be more for hired AV than for owned AVs but it all depends much on how they are used, as discussed in the analysis of round 1.

9a and b. What do you expect to be the impact on Urban bus Public Transport demand when AVs are 10% or 20% of the car fleet?

I made it clearer that this question refers to urban buses now. Overall the average expectation is a drop in bus demand of around 10% (for AV 10%) with a marginally greater one for 20% AV penetration. These figures hide regional variations. In the US & Canada, the reduction is closer to 30% where in the rest of the world (including ROW) the decline in demand is minimal, perhaps reflecting greater public transport mode share, coverage and frequency.
Pic 18.png
Looking at the chart, one can see that some respondents expect AVs to contribute to bus demand. The comments provide some suggestion for the variety of responses. It is argued that the use of hired AVs will reduce car ownership and therefore there will be many situations in which a bus will be almost as good as an AV and possibly cheaper. This argument, of course, is city dependent and even where in the city one resides. It is likely that low-frequency bus routes will suffer most as they are easily replaced by AVs. Therefore, it will be important to adapt these results to each local context.

10a and b. What do you expect to be the impact on Urban Fixed Track Public (Rail, metro, LRT, BRT) demand when AVs are 10% or 20% of the car fleet?

There is an agreement that fixed track, with protected right of way, will suffer a smaller reduction in demand than buses. This is partly because of the greater speed and insulation from congestion and partly because they are only justified when demand is high and therefore frequency, a defence against AV availability is also high.

Fixed track systems also seem to fit better with AVs as feeder modes, now that no large Park & Ride parking spaces are required. This explains why some respondents expected fixed track services to gain demand with AVs.
Pic 19.png
There remain many unknowns in establishing a sensible expectation for the impact of AVs on public transport.

11a,b and c. What would be the effect on the behavioural (for demand modelling) Value of Travel Time Savings (VTTS) on the part of AV users, given that they can legally undertake other activities while travelling in them? Express the effect as a ratio over the behavioural VTTS when driving a conventional car?

Driving is very demanding: holding a conversation with a passenger or listening to an inspirational recording, are allowable multi-tasking activities with driving. Texting and reading are not; and holding a hands-free telephone conversation although legal in many countries is recognised as a risky distraction.

Rail and some bus travellers can perform other tasks while travelling like working on a laptop or updating their Facebook profile. Even sleeping is a worthwhile “activity” on a flight for the overworked professional. If instead of focusing on driving the AV traveller can perform other useful activities the one would expect a reduction in the subjective value placed on shortening the journey; this would imply a reduction in behavioural VTTS.

The prevalent view from the respondents seem to be that there will be a 10% or so reduction in the behavioural VTTS, that is the parameter in the utility function or generalised cost that multiplies travel time. This is probably reasonable. One can observe that taxi passengers do not display behaviour consistent with a large reduction in behavioural VTTS (partly because many taxis charge by time and distance). But even VIPs driven in limousines at no charge do not seem to have a much reduced VTTS.

There is, however, a wide dispersion in the responses that makes me doubt whether we have all interpreted the question in the same way. As shown in the chart, a non-trivial number of respondents believe that the VTTS will increase when travelling on an AV. Looking at their comments there are some puzzles. Some argue that it will be rich people that would purchase and use AVs first and therefore will have a higher VTTS (but in my mind not necessarily higher than when they were travelling in a normal car); another argued that many people, especially commuting, would prefer not to do “something useful like work” but it is difficult to see why this would increase VTTS.

Although the impacts on different journey purposes vary, it is difficult to identify a consistent pattern in these variations.

For those who consider that the Value of Travel Time Savings could not possibly increase when using an Automated Vehicle the average value of the ratio AV_VTTS/Car_VTTS is about 0.80 (calculated by ignoring values above 1.0) for all the journey purposes.

Pic 20.png
Some respondents were quite aggressive assuming that the VTTS could be even halved under AV operation, not my personal view.

12a,b and c. What would be the effect on the Social Value of Time to be assigned to AV traveller (given that they can legally undertake other activities while travelling in them) in Cost-Benefit Analysis. Express the effect as a ratio over the Social VTTS when driving a conventional car?

This was not a modelling question but one that applies to project or plan appraisal. In most countries an “Equity” Value of Time is used in Cost-Benefit Analysis to contrast the time-saving benefits from a project with monetary costs and benefits. This is an equity value not directly extracted from behavioural data as doing so would bias projects in favour of higher income groups. Governments also try to avoid biasing projects in favour of a group or mode unless a policy decision dictates it.
On average, a similar reduction in Social VTTS to that on Behavioural VTTS was suggested by respondents. However, around 50% (depending on purpose) think that the Social VTTS should remain about the same as now.


It is difficult to imagine reasons why the social or equity VTTS would be greater in an AV than on a conventional car; nor that the reduction in Social VTTS could be as large as 50%.

From a government point of view, the easiest default assumption would be that the VTTS would be the same and any gain in productivity will be an upside to the country as a whole; perhaps not as valuable as the accident reduction power of AVs but worthwhile nevertheless.

For those who consider that the Social or Equity Value of Travel Time Savings could not increase when using an Automated Vehicle the average value of the ratio AV_VTTS/Car_VTTS is about 0.88 (calculated by ignoring values above 1.0) for all the journey purposes.

Additional questions

The Delphi Panel was also offered the opportunity to comment on factors influencing major choices in respect of CAVs, owned or hired.

What are, in your view and region, the most important Social or Cultural drivers that will affect the choice of owning or hiring AVs as MaaS in the future? For example, a cultural driver could be the idea that owning a car is only for men, not for women and therefore mostly women will hire AVs.

The most important cultural driver, mentioned by the vast majority of respondents, is the value attached to owning a visibly attractive piece of hardware. This was described variously as the “pride of ownership”, “self-realization through ownership”, “status symbol”; all promote ownership over hiring.

Second in importance among respondents was the type of use made of the vehicle that strengthens the case for ownership:
  • The need to customise it to carry children of different ages and requirements (child seats, emergency nappies, bicycles);
  • The need to store and carry tools of your trade: power and hand tools, tuba, drum kit;
  • The advantage of storing and carrying some leisure equipment like golf clubs, yoga mats, gym bag, etc.
  • The pursuit of outdoor activities would favour ownership or at least rental of an appropriate vehicle to support them when needed.
Age, or willingness to explore new technologies. It was suggested that younger people are less keen to own cars today (true in most advanced countries) and more willing to try new ideas, therefore they would be early adopters of AVs for hire. Older people may feel unsafe in an AV.

The quality of the service offered by hired AVs. If AVs turn up dirty, in poor shape and you suspect it may not be properly maintained then the attraction of hiring will be diminished significantly.

High urban density would favour hiring over ownership as it would take less time to secure an AV (as there would be more nearby).
High frequency of use, for example daily commuting or using it for work, would favour ownership.

Reverse urbanisation, flocking back to the downtown, would favour hiring.

Capital versus Operating Costs. This issue turns up under other headings as well. It is well known that most car owners do not perceive capital or operating costs very well. Capital is a sunk cost and does not affect much de decision to use the car. Vehicle Operating Costs (VOC) are rarely perceived in full; fuel costs alone are vaguely associated to travel whereas maintenance, insurance and license costs are just recurrent nuisances to the privilege of owning a car. In contrast, a hired vehicle will be priced to account for all of these, including depreciation (and AVs, as computers, may suffer higher technology depreciation). This issue of perceived costs will work against hiring AVs, at least for those already used to owning a car.

Multiple car families may opt to replace one car with AV MaaS.

In some regions where women are not allowed to drive alone, the jury is out about whether they would be allowed in an AV.
Large families would be expected to prefer ownership while small ones and singles would prefer hiring. An alternative version is: a large family may prefer to own one vehicle and hire or rent AVs to complement basic needs, for example, days out.

What Technical aspects of AVs will influence the decision to hire or own them and in what direction? For example, a very visible technology (an ugly LIDAR unit visible on top of the car) will make it more likely to hire them.

Here the main driver was the reliability of owned or hired AVs. This will depend, of course, on future regulations and their enforcement. Concerns about:
  • The state of maintenance of hired vs owned AVs
  • The up to date nature of the AV operating system, version control
  • Safety and liability concerns
The key question is whether hired AVs will be subject to more stringent and regular checks as they would be, in essence, public services (as are taxis and their drivers); and whether one can rely on owners to keep the technology up to date (how many OS updates are postponed?).

Fast change of technology, even by 2030 the focus of this survey, was seen as a factor favouring hiring AVs. If it was easy to upgrade an AV, downloading firmware, for example, this would support ownership; but the general impact would be to prefer hiring to avoid technical obsolescence.

Additional, but very practical concerns were about the cleanliness of the hired AVs and the possibility of personalising it with cloud-based data for seat adjustment, media player, temperature, etc.

Integration with Public Transport services (positive feature) was seen as more likely with hired rather than owned AVs (MaaS again).

The acceleration profile of AVs was of interest in some locations. If hired AVs are less agile than owned ones (hired Prius vs owned Tesla AVs) this may influence the choice to buy. On the other hand, if the ride is smoother than currently offered by aggressive car owners and taxi drivers, then hiring may prove attractive.

The security of personal information, protection from hacking it, was seen as very important but unclear whether this would be better or worse in hired vehicles as one may store more personal information in your own AV, or your AV cloud.

What Economic or Price issues are likely to influence the decision to hire or own AVs?

The issue of perceived versus real costs was central to this section. The tendency to hire would be greater the greater the understanding of the real costs of motoring. The temporal depreciation of vehicles was probably the most difficult issue to communicate. Out-of-pocket costs are always perceived more onerous than capital and annual costs.

The critical premium of an AV over a conventional car of similar specifications was identified at around US$5,000. If the premium falls below this figure the incentive to own AVs will increase significantly. Before that, AV users are probably more likely to hire them.

The other issue related to this is who pays for the empty movements of AVs. In fact, both owners and hirers will pay for the internal, but not the external costs of empty movements. Governments will find it easy to justify a Road User Charge (RUC) on zero passengers AVs to internalise its externalities (no victims). This is the ZVM or ZVK issue. It will then be “natural” to extend the RUC to AVs and other vehicles when travelling with passengers.

The price structure of hired AVs would also be of relevance. Some people may favour packages guaranteeing certain mileage per month, or a level of availability (maximum waiting time, like the level of contention in your broadband connection).

The opportunity to hire out (Airbnb like) your own AV when not in use will reduce the cost of ownership and create a nightmare for regulators. Ride-sharing, facilitated by the central management of AVs, may be a way to reduce costs and meet new people.

How AV hiring companies will handle surges will affect the attraction of hiring; the availability of sufficient spare fleet to cope with surges spreading the cost is often more attractive to users than price demand management.

Do you think there will be environmental influences or concerns that will influence the proportion of AVs hired or owned?

Most respondents thought that there were no clear environmental issues related to hiring rather than owning AVs. It was mostly an issue of the type of energy used and there was no reason to believe that hired AVs would be much more different from owned ones. This implies mostly overnight charging of electric vehicles will be sufficient for a 16-hour shift as they would be more intensely used than owned ones.

An important practical point was raised in respect of emissions and vehicle disposal costs. It was presumed that fleets of hired AVs would be renewed more often disposing of (for recycling), rather than reselling, the AVs after their economic (or environmental) life had expired. A younger fleet would be less polluting.

Concern was expressed that the ZVK problem will encourage RUCs and this was, by some, seen as unpopular; others thought that they could be introduced at last.

What type of transport policies are most likely to influence the decision to hire or own AVs?

Parking policy came top of the list by a large margin. Parking restraint and high parking charges would favour hiring. Good Park & Ride facilities would reduce the problems of owning AVs.

Different tax and charges regimes for privately purchased AVs, presumably higher than for fleet purchases, would favour hiring.

These two policy instruments seem quite useful to manage both the penetration rate of AVs and the proportion of hire/owned AVs. Governments may think it desirable to promote AVs to save lives and reduce congestion.

The price structure of any RUC may also affect hire/own proportions. For example, RUC may charge for the length of stay on on-street parking spaces.

Frequent and strict inspections would favour safety and hiring as owners dislike interruptions in the availability of their vehicles.

In the longer term, land use policies supporting densification and transport policies strengthening public transport would favour hiring AVs.

What legal aspects of using AVs are most likely to influence the decision to own or hire?

Liability is again top of the list. Who is liable when something goes wrong, the owner or the manufacturer; presumably not the hirer of the L5 AV. This implies that some authority validates the road-worthiness of the vehicle and how often (presumably more often for hired AVs).

Would you be allowed to send a young kid (less than 16 years old) on its own to school on your own AV but not on a hired AV?

This would also depend on the possibility of taking control of the vehicle; can we assume that all L5 AVs will not allow the user to take charge?

Is there any ethical consideration that is likely to influence decisions to hire or own AVs in the future?

The most frequent view was that hiring was seen as available to all whereas owning was less equitable. Equity and fairness are big issues today, c.f. Trump and Brexit. But this is not only an issue of income but also of adaptability to new technology; therefore, training will be needed not to leave people behind.

Tying up much wealth on hardware that is not fully used, when there are other demands, was seen as a similar ethical issue. A related issue was an evolution towards an ethical view rejecting consumerism/elitism and preferring services and goods more widely available to all.

Using AVs should be seen as the preferred ethical stance as they will be safer, life and suffering will be spared. Will self-driving be seen as the new smoking? This will not affect much the hiring ratio except that hiring would be easier than owning.

Replacing workers with machines and computers was also perceived as an ethical issue not too different from replacing local workers with foreign ones.
Children alone in AVs was seen as an ethical as well as a legal issue.

Concluding remarks

There was the hope to identify a few drivers for each topic and then run a second survey to try to allocate weights of importance to each.

As modellers, we know that context is paramount and interpretation unavoidable and necessary. Therefore, this list must be put into your local context, assess with others (not just colleagues) their relative importance and likely evolution to your planning horizon. Then, it will be important to identify objectives for any intervention and decide which levers are available and may work better locally. Then, track spontaneous factors (like pride of ownership), policy levers and the evolution of AVs, hired, rented and owned.

The figure below tries to summarise the implications of the seven main drivers for different segments of the population of London. This is only a desk-based example, not a real one. It attempts to visualise the most important drivers within each group and imagine how they would influence four different segments: singles, small family (up to 3 members), large family and over 65. The figure assigns a rating between 0 and 100, where 100 is maximum tendency to hire rather than share.