Self Driving Cars - The Complicated Questions
As the prospect of self driving cars looms closer, complicated questions are being asked. AI being behind the wheel places humanity in uncharted waters with tough political and moral questions.
Will self driving vehicles put 1000s out of a job? During an accident, who takes priority the owner or amount of lives? In situations where instinctual and heat of the moment decisions would be made, now have to be intentional calculated decisions. In grey areas on the road, how will AI be able to read the situation of human drivers without human communication?
Self driving cars will certainly shape a new era for the roads, but there are tough obstacles in the way. Today I will take you through some of the uncertainties around this technology.
Will self driving vehicles put 1000s out of a job? During an accident, who takes priority the owner or amount of lives? In situations where instinctual and heat of the moment decisions would be made, now have to be intentional calculated decisions. In grey areas on the road, how will AI be able to read the situation of human drivers without human communication?
Self driving cars will certainly shape a new era for the roads, but there are tough obstacles in the way. Today I will take you through some of the uncertainties around this technology.
Driverless roads, a jobless road?
The automotive industry as a whole is certainly going to change, not decline, but it’s the manual driving jobs that will likely lead to an inevitable demise.
Self driving lorries are currently being tested and there will be a day in the future where self driving lorries are rolled out on a mass scale by big companies. Is it something to be worried about now? I don’t think so. There are many complicated issues that are still unanswered.
Self driving lorries are currently being tested and there will be a day in the future where self driving lorries are rolled out on a mass scale by big companies. Is it something to be worried about now? I don’t think so. There are many complicated issues that are still unanswered.
The Critical Decision: Who Lives?
In the event of a hazard, the tyre bursts and the car has lost control, a critical event plays out.
With a human being, an instinctive or unconscious decision would be made in that split second. AI works differently in decision making. Absolutely everything has to be a calculated decision. AI essentially has to be programmed a rule set of actions to take. This creates a whole set of complicated political choices the creators must decide upon.
Your car tyre bursts and the car is hurdling out of control. The car has 3 choices.
Protect the owner at all costs? Protect the number of lives at all costs? Or even go with the decision that would cost the least amount of financial damage?
In theory, the AI could be programmed to have freedom to choose at the time, but leaving the trust in AI does not sit well with a large number of debaters.
With a human being, an instinctive or unconscious decision would be made in that split second. AI works differently in decision making. Absolutely everything has to be a calculated decision. AI essentially has to be programmed a rule set of actions to take. This creates a whole set of complicated political choices the creators must decide upon.
Your car tyre bursts and the car is hurdling out of control. The car has 3 choices.
- Crash into the wall resulting in potential fatality for the owner/ passengers of the car.
- Hurdle into a crowd of people resulting in a grim fate for a group of pedestrians but protecting the passengers/owner of the car.
- Or crash into traffic resulting in potential fatality to numerous other drivers but possibly protecting the passengers/owner of your car.
Protect the owner at all costs? Protect the number of lives at all costs? Or even go with the decision that would cost the least amount of financial damage?
In theory, the AI could be programmed to have freedom to choose at the time, but leaving the trust in AI does not sit well with a large number of debaters.
A Bright but Hazy future?
The prospect of self driving cars has many great benefits and this isn’t all doom and gloom. It will likely create competition for Taxi’s leading to a much cheaper price for the consumer. This is great for tourists and people who can’t afford a car. It could reduce the amount of drink driving incidents. With no human error, it should lead to far less accidents on the road.
However, a big problem lies with other human beings and AI being on the road, mix that with the current political roadblocks and this leaves a lot of uncertainty in the air.
How will other humans react to this? Not just in general opinion and attitude but in practical situations on the road? How can AI react to human error, if there’s still human drivers on the road, it presents another hurdle for AI due to the unpredictable nature of human beings.
However, a big problem lies with other human beings and AI being on the road, mix that with the current political roadblocks and this leaves a lot of uncertainty in the air.
How will other humans react to this? Not just in general opinion and attitude but in practical situations on the road? How can AI react to human error, if there’s still human drivers on the road, it presents another hurdle for AI due to the unpredictable nature of human beings.
Can AI Adopt Human Interpretation?
In certain grey areas on the road, human drivers can interact with each other through looking at the driver. A good example is at a mini round-about, when all of you are at the intersection, everyone at this point has to give way to their right. Drivers read the situation through understanding who will take the lead in the situation via either looking at the drivers or anticipating someone’s move. This is something that self driving cars may fail to interpret.
Human drivers also make many minor mistakes on the road in which currently driverless cars struggle to understand.
In terms of pedestrian behaviour, there are other issues that lie. According to computer ethics commentator Ben Byford , both understanding how the AI will behave and not knowing both pose problems. Human beings have a somewhat understanding of how another human will react which makes you feel safer crossing the road. With many car manufacturers developing their own systems, the behaviour could become fragmented.
“If all the major car manufacturers have a different version of this system, it's going to make no sense whatsoever. And things can snowball into a catastrophe when people don’t understand how these things work”
However, the problem is if we did understand the AI’s behaviour, it is likely to be of a predictable linear nature meaning situations can easily be manipulated.
“Say that the cars always react in the same way. That being the case, I could seriously injure people by walking out in front of a car on purpose, knowing that it wouldn't be able to hit me.”
In a nutshell, the AI will need to fully understand human behaviour as well have the intuitiveness to adapt to unique situations. As well as that, human beings will need to understand how self driving cars behave, which in grey areas without the ability to communicate or give off facial signals will be some challenge.
I think there are a lot of huge challenges for self driving cars and really are only feasible currently when no other humans are present. Self driving cars seems inevitable but not imminent. The real question is when will self driving cars become a standard? I really don’t think anytime soon.
In terms of pedestrian behaviour, there are other issues that lie. According to computer ethics commentator Ben Byford , both understanding how the AI will behave and not knowing both pose problems. Human beings have a somewhat understanding of how another human will react which makes you feel safer crossing the road. With many car manufacturers developing their own systems, the behaviour could become fragmented.
“If all the major car manufacturers have a different version of this system, it's going to make no sense whatsoever. And things can snowball into a catastrophe when people don’t understand how these things work”
However, the problem is if we did understand the AI’s behaviour, it is likely to be of a predictable linear nature meaning situations can easily be manipulated.
“Say that the cars always react in the same way. That being the case, I could seriously injure people by walking out in front of a car on purpose, knowing that it wouldn't be able to hit me.”
In a nutshell, the AI will need to fully understand human behaviour as well have the intuitiveness to adapt to unique situations. As well as that, human beings will need to understand how self driving cars behave, which in grey areas without the ability to communicate or give off facial signals will be some challenge.
I think there are a lot of huge challenges for self driving cars and really are only feasible currently when no other humans are present. Self driving cars seems inevitable but not imminent. The real question is when will self driving cars become a standard? I really don’t think anytime soon.