Self-Driving Cars and the Law
Introduction
The idea of self-driving cars is always as old as the car itself. In
1939, Futurama General Motors posed the idea. However, in
the 1990s, the thought became more serious. Platooning, a
Dutch company, gave a demo in 1998. In 2021, semi-
autonomous cars have been introduced on the market. All
major car manufacturers, technology suppliers (e.g., Samsung)
and tech companies such as Google, are working on self-driving
cars.
The expectations are high and the benefits that self-driving
cars would bring are substantial. For example:
o Safety – Human error is involved in more than 90% of
road accidents. If we can remove the human error factor,
less road accidents should occur.
o Road capacity – Harmonisation of traffic flow can lead to
increasing the capacity of existing roads.
o Driving comfort – Use 0f travel time may be used to
complete other tasks instead. Those who are nervous
drivers may have their anxieties reduced.
o Transport cost savings
o Offering mobility to more people
Legislative action
Governments are trying to facilitate the developments by
taking legislative action. Legislative action in many countries
allows the testing of self-driving cars on the public. Countries
such as Germany and the UK (Automated and Electric Vehicle
Act 2018) have started to revise their legal frameworks to pave
the way for automated driving.
So, will self-driving cars be introduced soon (say, within 10
years)? To get a better understanding of the answer to this
question, it is useful to refer to the levels of automation. There
are 5 different levels of automation, with level 1 meaning no
automation and level 5 meaning full automation (i.e.,
completely self-driving car):
1. Driver is partly assisted; cruise control or lane keeping
assistance.
, 2. Hands off / feet off driving possible, but driver must
continuously monitor system performance and be
prepared to intervene; auto-pilot modes.
3. Driver may turn attention away from traffic/vehicle control
(e.g., read a book, text etc.) but must be prepared to
intervene when needed; current test vehicles.
4. As level 3, but system can autonomously revert to
‘minimum risk condition’ if needed; highway driving pilot
mode.
5. No driver needed.
There is an essential step from level 2 (partial automation) to
level 3 (conditional automation). The current available ‘auto
pilots’ are considered level 2 and are far from perfect. Level 3 is
tricky to achieve. The request to intervene must be ‘timely’ and
human factors research shows that once the human is
distracted, situational awareness is lost, and it can take them
time to adequately respond to the system if needed.
Man vs. Machine
Technology
- Reacts much quicker.
- Doesn’t get tired, distracted, or intoxicated.
- Can see more than a human driver (360 degrees).
- May ‘communicate’ with other vehicles and/or the
infrastructure.
Humans
- Still better at making sense of the complex and dynamic
driving environment.
- Can interpret the behaviour of other road users.
- Can account for diverse weather and road conditions (poor
road marking may confuse a self-driving car, for example).
- Can react to unexpected things you might encounter in
daily life whilst driving. It will take a very long time to train
AI to encounter every possible thing that a human might
from day to day.
Real-life incidents
The first real life incidents with semi-autonomous vehicles
ended in severe crashes. For example, Tesla’s autopilot mode
Introduction
The idea of self-driving cars is always as old as the car itself. In
1939, Futurama General Motors posed the idea. However, in
the 1990s, the thought became more serious. Platooning, a
Dutch company, gave a demo in 1998. In 2021, semi-
autonomous cars have been introduced on the market. All
major car manufacturers, technology suppliers (e.g., Samsung)
and tech companies such as Google, are working on self-driving
cars.
The expectations are high and the benefits that self-driving
cars would bring are substantial. For example:
o Safety – Human error is involved in more than 90% of
road accidents. If we can remove the human error factor,
less road accidents should occur.
o Road capacity – Harmonisation of traffic flow can lead to
increasing the capacity of existing roads.
o Driving comfort – Use 0f travel time may be used to
complete other tasks instead. Those who are nervous
drivers may have their anxieties reduced.
o Transport cost savings
o Offering mobility to more people
Legislative action
Governments are trying to facilitate the developments by
taking legislative action. Legislative action in many countries
allows the testing of self-driving cars on the public. Countries
such as Germany and the UK (Automated and Electric Vehicle
Act 2018) have started to revise their legal frameworks to pave
the way for automated driving.
So, will self-driving cars be introduced soon (say, within 10
years)? To get a better understanding of the answer to this
question, it is useful to refer to the levels of automation. There
are 5 different levels of automation, with level 1 meaning no
automation and level 5 meaning full automation (i.e.,
completely self-driving car):
1. Driver is partly assisted; cruise control or lane keeping
assistance.
, 2. Hands off / feet off driving possible, but driver must
continuously monitor system performance and be
prepared to intervene; auto-pilot modes.
3. Driver may turn attention away from traffic/vehicle control
(e.g., read a book, text etc.) but must be prepared to
intervene when needed; current test vehicles.
4. As level 3, but system can autonomously revert to
‘minimum risk condition’ if needed; highway driving pilot
mode.
5. No driver needed.
There is an essential step from level 2 (partial automation) to
level 3 (conditional automation). The current available ‘auto
pilots’ are considered level 2 and are far from perfect. Level 3 is
tricky to achieve. The request to intervene must be ‘timely’ and
human factors research shows that once the human is
distracted, situational awareness is lost, and it can take them
time to adequately respond to the system if needed.
Man vs. Machine
Technology
- Reacts much quicker.
- Doesn’t get tired, distracted, or intoxicated.
- Can see more than a human driver (360 degrees).
- May ‘communicate’ with other vehicles and/or the
infrastructure.
Humans
- Still better at making sense of the complex and dynamic
driving environment.
- Can interpret the behaviour of other road users.
- Can account for diverse weather and road conditions (poor
road marking may confuse a self-driving car, for example).
- Can react to unexpected things you might encounter in
daily life whilst driving. It will take a very long time to train
AI to encounter every possible thing that a human might
from day to day.
Real-life incidents
The first real life incidents with semi-autonomous vehicles
ended in severe crashes. For example, Tesla’s autopilot mode
