Month: October 2018

The trolley problem is a big con

In 2016 the Massachusetts Institute of Technology launched a global online survey to test whether there are cultural differences in response to the trolley problem, applied to robot cars.

Packaged as The Moral Machine, it asked players to instruct robot cars to “choose the lesser of two evils, such as killing two passengers or five pedestrians”.

This type of exercise is a pernicious con orchestrated for the benefit of the car manufacturers. Here is why.

The trolley problem hardly ever comes up in real life

If one looks at real life situations, where car drivers kill or injure pedestrians or cyclists, it is extremely rare that the driver was “choosing the lesser evil”, i.e. was trying to save another life. The trolley problem is non-problem, a masturbation for so-called theoretical ethics professors.

The trolley problem is an attempt to frame the ethical issues involved with robot cars for the benefit of the machines

By focusing on the trolley problem, i.e.a non-problem, the industry is perversely trying to shape public opinion towards its objectives.

It promotes the fallacy that robot cars will kill only to save lives

Robot cars are promoted as being much safer than human operators, capable of erasing the million+ yearly butchery on our roads. It is trying to sell the idea that robot cars will kill only when forced into an impossible dilemma. The truth is that robot cars have and will continue to kill for very different reasons, namely

  • the inadequacy of the software to deal with anything more sophisticated than the cultural desert of a North American suburb.
  • the vulnerability of robot cars to hacking attacks
  • the willing interference by operators of safety systems to improve performance, as when evil Uber switched off one of the braking systems in a car that killed a pedestrian.

It does not focus on the real issue: why would the robot car be in such a position to be faced with the trolley problem

The key moral Vision Zero question is “what measures do we need to take to ensure that robot cars do not kill innocent citizens?” In other words we need to treat robot cars as fallible machines, and we need to create environments where they are very unlikely to kill or seriously injure. this means adopting similar measures that Vision Zero mandates for human-driven machines:

  • minimise their presence in busy urban areas
  • limit their speed, to reduce the impact of crashes (much easier to enforce with robot cars)
  • devote road space to bicycle-only travel

You can understand how car manufacturers can be against the Vision Zero agenda, which necessarily limits the freedom to drive as and where one pleases.

It is a pre-emptive fight to disable citizens from being able to reduce the mobility of robot cars

Car manufacturers are terrified by the idea that if safety guidelines are too strict, and robot cars are regulated in such a way to prevent all avoidable crashes, the cars would not be able to move because anyone can step off the pavement and block the progress of a robot car. A Vision Zero robot car would not be able to move. The car manufacturers are always lobbying to prevent any interference with the privileged position of drivers and are already trying to make sure that such a scenario will not be contemplated by regulators.

It is a victim blaming exercise

By framing the issue as a trolley problem, they are shifting the blame from the cars to the pedestrians. Note how the MIT survey tries to show the cultural differences to whether a robot car should kill someone crossing the street not at a designated crossing site, (Global preferences for who to save in self-driving car crashes revealed). It perpetuates the immoral proposition, so dear to the British (un)Justice System, that drivers are allowed to act as vigilantes, absolved of blame if they kill a pedestrian who has broken the rules, rules of course that have been designed prioritising the convenience of drivers over the safety of ordinary citizens.

We are not against the idea of robot cars

A technology that has the potential of reducing the consequences of human error is to be welcomed. However, we have to be very careful that the car industry is hi-jacking what can be a life-saving technology, with the goal of entrenching the privileges of its customers. The safest car is a car off the road. The reduction of the number and the speed of vehicles on our roads is a much more effective way to eliminate avoidable road deaths. That should the primary goal of urban planners. Opposed to this vision of cities designed for the benefit of ordinary citizens, robot cars are being sold as the solution of future mobility: that is snake oil and the trolley problem is a pernicious way to frame the public discourse for the benefit of car manufacturers.

The proper response to people pushing the trolley problem is “Wank off”.

Finally, here is an anecdote that shows the total lack of ethics of many involved in pushing robot cars:

 

Advertisements

City of London to adopt Vision Zero

The City of London’s draft Transport Strategy is an ambitious plan which puts walking and cycling at its core. Integral to the strategy is the adoption of Vision Zero concepts:

We will deliver Vision Zero to eliminate death and serious injuries on the City’s streets by 2040.
Measures to deliver Vision Zero and reduce road danger will be delivered across four themes:
• Safer streets
• Safer speeds
• Safer vehicles
• Safer behaviours
This means:
Being proportional in our efforts to tackle the sources of road danger, focussing on those users of our streets who have the greatest potential to harm others due to the size and speed of their vehicle
• Recognising that people will always make mistakes and that collisions can never be
entirely eliminated. Our streets must therefore be designed, managed and used to cater for an element of human error and unpredictability
Reducing vehicle speeds on our streets to minimise the energy involved in collisions and protect people from death or injury
Seeking to reduce slight injuries and fear of road danger alongside the principal focus on eliminating death and serious injuries

Here are some details.

Safer streetsSeven dangerous streets/junctions (including St Paul gyratory, High Holborn and Aldersgate) will be redesigned by 2030.  They will also be narrowing and raising the entrances to side streets to require drivers and riders to manoeuvre more slowly
Safer speeds – Adoption of a City-wide 25kph speed limit by 2022
Safer vehicles – Improving the FORS standards and widening the scheme to coaches and vans
Safer behaviours – Among various measures, “Encouraging TfL to require safety training as part of private hire and taxi licensing. This will include Bikeability Level 3 training”

In order to ensure that the Strategy translates in real action, the City proposes a Road Danger Reduction Action Plan, “a five-year delivery plan for measures to achieve Vision Zero and implement the Safe Systems approach”.

Overall the Strategy shows ambition. It also acknowledges that there is a lot of work to do:

Only 4% of people currently consider the experience of cycling in the City to be pleasant (and 56% consider it to be unpleasant). We want this figure to be 75% by 2044. More than half of people cycling in the City scored their feeling of safety while cycling as a 1 or 2 out of 5.

On average 19 people cycling have been killed or seriously injured on our streets every year for the last 5 years

Pedestrian deaths and injuries per billion journeys by foot

Rachel Aldred has divided the casualty data by the average distance people walk and the results are quite startling:

Someone walking in Barking runs more than twice the risk of being seriously injured or killed than if she were walking in Kingston.

Both boroughs have low walking shares, so is the difference linked to income levels?

Dr. Aldred will conduct more research.

Meanwhile, here is a couple of possible explanation:

  • Kingston’s affluent residents demand money to be spent on a beautiful public realm; that often means better and safer infrastructure for pedestrians; Kingston was also recipient of Mini Holland money, and safe cycling infrastructure also improves the safety of pedestrians. On the other hand Barking, where Ford used to have a large factory, is still married to “car culture” and its authority is not devoting sufficient attention to pedestrian safety.
  • However, there may be another factor that may influence the income/danger correlation: culture. In poor neighbourhoods, a car is still perceived as a status sign; car owners are more likely to act disrespectfully towards non-owners. Additionally, poor neighbourhoods have a larger percentage of immigrants from countries (such as India, Arab and African countries) where this disrespect is endemic. In more affluent neighbourhoods, people don’t need to prove their status by driving aggressively.