Driverless Vehicles, a Techno-Utopian Dream Sold As a Public Good
Automation should complement the human agency and not replace it
As negative side effects of automation become more evident, the world is realising that for many kinds of work we don’t want to replace human beings, even if the work is a little dull, dirty or dangerous. In fact, that replacement can become counter-productive. (Gill Pratt, CEO of Toyota Research Institute)
This post is not about the intricacies of the Autonomous Vehicle (AV) technology because I am a generalist with little technical knowledge. I only look at the philosophy of the AV technology, the folly of some of its underlying assumptions, and its impact on society and human autonomy.
“We will see self-driving cars on the roads next year!” We have been hearing this statement for the last ten years. As Dr James Kuffner of the Toyota Research Institute said, “Let me be clear: nobody is even close.”.
The self-driven vehicle is an omnibus term. There are six levels of vehicle autonomy.
The U.S. Department of Transportation has adopted the six levels of driving automation as defined by The Society of Automotive Engineers (SAE).
SAE defines 6 levels of driving automation ranging from 0 (fully manual) to 5 (fully autonomous).
A fully autonomous vehicle comes under L-5, where no human attention or interaction is required.
We will not see any L-5 vehicle on public roads in the foreseeable future except in dedicated corridors which will disqualify its fully autonomous status because such a protected environment is not all-weather terrain.
Modern sophisticated technologies possess immense seductive power. They capture our imagination about their possibilities and make us suspend our disbelief.
Creators of new technologies are clever. They hitch their technologies to the public interest bandwagon. They have projected the AV technology as a panacea for the following chronic problems:
- Driverless vehicles will save lives. Every year, over 40, 000 Americans die in road accidents caused by stupid and careless human drivers.
- Self-driving vehicles will decongest the roads as people will abandon car ownership for ride-sharing.
- AV technology will bring down pollution levels and reduce carbon emissions.
Let’s examine these claims.
Road safety
An article in Forbes.com quotes research conducted by the Insurance Institute for Highway Safety (IIHS) which says self-driven cars won’t prevent most accidents.
The IIHS studied the data for 5000 road crashes and warned that the technology likely will only prevent one-third of all vehicle crashes.
The AV technology can avoid typical human errors like people driving under the influence of alcohol and drivers falling asleep.
What the technology can’t prevent are what the institute calls “prediction errors,” such as misjudging the speed of other vehicles, excessive speed for treacherous road conditions and mistaken driver efforts to avoid a crash.
In 2018, an Uber autonomous vehicle killed a woman who was walking her bicycle across a major thoroughfare. The vehicle failed to swerve out of the way despite the presence of a human in the driver’s seat.
Can an L-5 car operate in mixed traffic in the company of vehicles driven by humans? Can it anticipate all the actions of human drivers?
Systemic errors are deadlier than human failures, especially when there is no human back -up.
We have had two deadly crashes of the Boeing Max 737 aircraft during the past two years, killing 346 people. In both cases, a system intended to prevent stalling malfunctioned. The pilots did not understand the systemic failure. Nor were they trained to rectify the error.
Even in an L-4 car, the human back up will struggle to take control when systems fail suddenly. The AV technology will struggle to adjust to sudden changes in weather, no matter how many scenarios they teach to the algorithms.
Hackers can attack the systems and cause accidents.
Congestion
A reduction in road congestion presupposes that people will abandon car ownership and will opt for ride-sharing. This rosy scenario is wild speculation. If not a pipe dream. It presupposes L-5 vehicles will take over the roads. The rosy future when a driverless car picks us up from our doorstep and takes us to our destinations is very far away.
Congestion will continue in mixed traffic because of a lack of coordination between L-5 vehicles and human-driven vehicles.
Even in corridors dedicated to fully autonomous vehicles, congestion will happen unless there is perfect coordination between the vehicles.
Pollution
Pollution will decrease if the entire AV fleet is electric. There are two conditions. The government will have to ban all non-electric vehicles on the road. The net reduction in carbon emissions will depend on whether they have sourced the electricity used to power the vehicles from renewable sources.
Full automation of driving is neither desirable nor workable
Knowledge is of two kinds. One is tacit knowledge. It refers to the stuff we do without thinking about it, like riding a bike, driving a car or reading a book. Surgeons use their tacit knowledge to perform surgeries. We embed this cognition and automatically retrieve it when we need it. The mental processing happens without our awareness.
Explicit knowledge is the other type. This is the stuff we can describe in writing, like how to change a flat tyre or how to solve a math equation. We can break down the process in steps and explain these steps to another person.
Driving is tacit knowledge. Despite the advances in machine learning, tacit knowledge will elude the grasping power of algorithms.
Conclusion
The social and economic problems caused or exacerbated by automation aren’t going to be solved by throwing more software at them. Our inanimate slaves aren’t going to chauffeur us to a utopia of comfort and harmony.. If the problems are to be solved, or at least attenuated, the public will need to grapple with them in their full complexity. To ensure society’s well-being in the future, we may need to place limits on automation. We may have to shift our view of progress, putting the emphasis on social and personal flourishing rather than technological advancement. (Nicholas Carr, in “Glass Cage: Who Needs Humanity Anyway?”)
Mindless automation that strips human agency of its dignity and value is neither desirable nor cost-effective. The social and economic costs of depriving millions of drivers of their livelihood are mindboggling.
Technology may not always create more jobs than it takes away. Where will the tens of thousands of truck drivers go, or how will they survive once the autonomous trucks take away their jobs? Write software for the computer systems onboard?
Fully automated vehicle technology is a techno-utopian dream if we consider its mass applicability. At best, fully autonomous vehicles only can operate in protected environments and in dedicated corridors. They can never take over transportation as predicted by the techno-utopians.
A better alternative to fantasies like fully automated transportation will be to develop the infrastructure for mass public transit in cities.
Will you fly in an aircraft without a human pilot merely because the cockpit is fully automated? Which of the following will you choose?: A free heart surgery performed by a sophisticated robot? Or one performed by an expert surgeon who charges a high fee?
Automation should be human-centric. It should complement the human agency and not replace it. We should not lose our ability to interrogate the techno- czars even if they espouse public interest to push through disruptive technologies with a high human cost.
Thanks for reading!
