In my last blog, I discussed the market growth of driverless cars, and identified a number of leading companies investing globally in this market segment, including: Alphabet, Amazon, Baidu, Google, etc. This blog focuses on driverless cars and AI ethics linking diverse sources to help guide board directors and CEO’s in ensuring they are on the right side of AI ethics vs the wrong side of the road.
There are many complexities in developing all the enablements to support a driverless car ecosystem, including balancing sensor-based hardware and AI software in outdoor environments that are under constant change. In addition, surface logistics, new regulatory and monitoring controls, and of course AI software ethical guidelines to govern the operational risks and decisions that a driverless car would need to navigate, if no humans were on board.
One of the most validated research surveys of machine ethics1, called the Moral Machine, was published in Nature, found that many of the moral principles that guide a driver’s decisions vary by country. This reinforces a regulatory set of complexities to navigate on cultural preferences. For example, in a scenario in which some combination of pedestrians and passengers will die in a collision, people from relatively prosperous countries with strong institutions were less likely to spare a pedestrian who stepped into traffic illegally. There were over 13 scenarios identified where a death could not be avoided and respondents had to make a decision on the impacts to old, rich, poor, more people or less people being killed. The research found that there are cultural variances in public preferences which governments and self-driving cars would need to take into account to gain varying jurisdictional confidence.
In other words different rules for countries would need to apply. Talk about moral and ethical complexities in design engineering. If this , then this etc.
Artificial Intelligence underpins self-driving cars, as diverse technologies are assembled, ranging from image recognition systems, using high powered smarter cameras, inside and outside cars, GPS (location sensors), voice and sound recognition systems, neural networks, high powered sensors, to make sense of complex patterns (ie: images from cameras, analyzing and recognizing cars and passengers in near by vehicles, trees, road signs, road movements, wind movements, in other words all environmental stimuli (noises/sounds/weather, etc.) brake movements, speed, etc), to assemble frameworks that can drive self-sufficiently. All of these technical inputs create an intelligent layer of insights and patterns that help self-driving cars operate efficiently but also allow broader operational dashboards across full network intelligence (like air traffic monitoring control systems) -all smart cars would be connected to a regulatory grid creating a much deeper connected infrastructure.
The more data that is collected by different ecosystem players present ethical challenges on
How a driverless car uses AI at Google (An Illustration). Reference:
MORE FOR YOU
- The driver sets a destination. The vehicle’s AI software predicts and ascertains a course.
- A turning, rooftop-mounted Lidar sensor screens a 60-meter range around the vehicle and makes a dynamic three-dimensional (3D) guide of the vehicle’s present environment.
- A sensor on the left back tire screens sideways development to identify the vehicle’s position comparative with the 3D guide.
- Radar frameworks toward the front and back bumpers ascertain distances to obstacles.
- Artificial intelligence programming in the vehicle is associated with every one of the sensors and gathers data from Google Street View and camcorders inside the vehicle.
- The AI recreates human perceptual and dynamic cycles utilizing deep learning algorithms and controls activities in driver control frameworks, like steering and brakes
- The vehicle’s software counsels Google Maps for early notification of things like tourist spots, traffic signs and lights and other obstacles
- An override function is accessible to enable a human to take responsibility for the vehicle.
Each of these experiences is using a foundational AI neural network with complex algorithms and machine learning methods to navigate the driverless car to its desired and safe location.
What are the Ethical Questions that a Board Director or a CEO should think about in relationship to driverless cars?
It is difficult to give justice to this type of question, but here are 5 key ethical pillars to start with.
1.Privacy and Security – Who will own the data that is collected? How much data will be collected in a driver less car, and can an end user have access to all the data and control any settings for their driver less car? How will the data patterns be used ie; shared with insurance companies or kept private?
2.Explainability and Transparency – Will the different data types be cleared explained and how these data types can be used so end users understand the risks and implications? Who is responsible for educating the end users/or companies that purchase driverless cars? What will be the role of the regulators ensuring explainability and transparency?
3.Accountability – Who is the ultimate owner/of the engineering underpinnings of the driverless car – the automobile manufacturer? In accidents, where there are conditions in roadways, construction conditions that confuse the car’s image recognition database, and accidents occur, who is accountable? What will the regulators do to create standard regulations?
4.Data Bias – Who will set the rules governing accident decisions on who will be killed with certainty or be spared and how will we ensure the public are supportive of the inherent decisions that will be made by a ingesting decisions made by man but enacted by a machine. Will countries vary who is spared in terms of a child, a man or a woman, or take into account their jobs, incomes and status in society? What role will equity and diversity play?
5.Sustainability – as there is high energy consumption requirements moving to driverless cars, what are the environmental risks vs other driving methods and most importantly the millions of jobs that will be lost when driver less cars are mainstream everywhere? What are the responsibilities of companies to transition and retrain their workforces? What are the macro implications of dis-intermediating entire logistics and transportation companies that have not modernized their fleets?
What is clear is all the rules and responsibilities will evolve and in many areas change. The meaning of responsibility or accountability will take on a major focus in enabling driverless cars to go mainstream.
Many new questions will need to be sorted out as outlined in Self Driving Vehicles, an Ethical Review, published in August 2021, an excellent review of the ethical challenges and call to action for more social research:
- If there is no driver who controls the vehicle, who is then responsible for the safety of its passengers and of those who travel or walk on the same roads?
- If a car is “driven” by a computer possessing artificial intelligence, does that intelligence constitute an entity that can be held responsible?
- What are the responsibilities of the vehicle’s current owner? Its manufacturer? The owner and manager of the road system? The organization running the traffic control centre that the vehicle communicates with?
The introduction of self-driving cars will challenge all of us to think harder about a large number of ethical issues that go beyond the common, extremely narrow, and will need to focus more on improbable dilemma-like scenarios such as network cyber-hacking across an entire automobile manufacturer fleet and replace safety controls with nefarious and dangerous settings or causing major power surges to disrupt supply chains. The outlier cases will help us make more informed decisions as it only takes one tragedy that was not accounted for as decision makers took the road most travelled by vs less travelled by. Its time to THINK very hard to get this right.
In summary, safety will always need to be the guiding ethical principle with clear responsibilities of different stakeholders defined clearly – front and center. The ethical considerations of driverless cars will impact the entire value chain, and will require integrated policy and legislative discussions that are inclusive of autonomous vehicle manufacturers, technology visionaries, law makers, regulator, distributors, infrastructure providers and government and public stakeholders. There are very complex ethical issues that need to be carefully thought through, in particular equity, privacy and risk implications.
Nature. The Moral Experiment.
MIT is offering a course on Smart Mobility: Reimagining the Future of Transportation Tech & Sustainable Cities