This post has been written with equal contribution by the authors of éclairé, Jai Kotia and Rishika Bharti.
It is a bright sunny morning and you are headed out for work. You make your way down the pavement and enter the location of your work place in an app on your phone. You hear a familiar sound coming from the direction of your car. The engine has started up, and the car is ready to begin the journey. You strap your seat belt on and ask the car to begin driving. The car makes its way onto the road and you are greeted by the sight of your neighbour. Except, he is sleeping in the back seat as his car is driving on. You soon reach the highway where every car seems to move in sync. You can’t hear any honks, even amongst hundreds of cars because they don’t need to. The sensors automate every movement of the cars. Its 2030 and every car around is self driven.
Excuse us for that Black Mirror-esque introduction. But whether you like it or not, self driving cars are soon going to take over the roads. In fact the cars look ready to hit the roads, it is just a matter of time before they get approval. See this video if you are still uncertain about how far self driving cars have come. In the light of this progress, it is important to understand and enforce a set of rules for self-driving cars, similar to the rules enforced upon human drivers. But it is a very complex process defining rules and discussing ethics for something that doesn’t have its own real life identity. Let us take a look at the AI ethics issues that we face in regards to self-driving cars.
Talking Points for AI Ethics in Self-Driving Cars
Who is to blame for the accidents?
This is a question that can grow into one of the most controversial subjects in world politics. Self-driving cars can also meet with accidents at some point. When they do, who will take responsibility for the loss? Will it be the company who built the car or the person who owns the car? What sort of punishment can be given to someone who has not contributed to the accident, since it was the AI controlling it? We feel that unless this important question is answered, it will be hard for self-driving cars to swing into action.
Security concerns over software hacking
Security of self-driving cars is threatened because software is prone to being hacked. It can in fact be a growing cause of kidnappings. Nefarious hackers can potentially hack into these self-driving car systems and hijack its passengers to a desired location. They can even steer the cars into accidents.
The personal information of the passengers is also compromised. If the system is hijacked, personal information such as credit card details and residential address is at a risk and can be easily intercepted via Bluetooth like technologies. These are pressing security concerns and if such a case occurs even once, it can quickly fear away people from use of such technology.
Rising pollution
Autonomous cars will make travel more convenient as they will allow everyone to locomote, especially those who were limited by the inability to drive themselves. Travel will be easier in general. As a result, number of miles travelled is likely to increase overall. If the engine is powered by gasoline, pollution will increase as a consequence. Thus before self-driving cars hit the road, it should be ensured they run on eco-friendly fuel variants.
Large scale unemployment of drivers
Yeah so your future Uber ride is going to be weird, because you won’t be making calls to your driver who is late anymore. There will in fact be no one you will be communicating with. Just an unoccupied car that will pick you up and drop you. It will play all your music and movies without any complaints and will gently ask for a 5 star rating through its human-like AI generated voice. From an economic standpoint, we are looking at a really large scale of unemployment. Millions of people all over the world are drivers by occupation. Only a few thousand people will be required to build these self-driving cars that will replace millions of jobs.
We might think that meeting halfway, by allowing the merger of automated cars with human driven ones could solve this problem. But this is not a fotersafe solution. Most of the prototypes that are in construction today are designed to avoid accidents. They may not be suitable for tackling unpredictable human-driven scenarios. An evidence of this fact is the Tesla Model S crash in 2015 where Joshua Brown was killed when Model S operating in autopilot mode crashed into his tractor-trailer.
While many argue that AI itself keeps generating jobs due to its increasing demand in all industries, they are very limited in number compared to the number of jobs they displace. They very aim of AI is to automate tasks without the requirement of human intervention.
Lack of regulation
There are no basic safety guidelines that are yet established to ensure the safety of self-driving cars. This is good for automobile companies as it makes it easier for them to put out cars but very harmful to the consumers. The industry is somewhat unregulated at this stage. A governing body needs to be set up to curate and enforce a set of regulations and safety standards. Futhermore, safety guidelines are extremely essential as self-driving cars are not as safe as they claim to be. Researchers speculate two main reasons for this:
- In times of crisis on the road, a human drivers response time is much less than that of the system installed in autonomous vehicles.
- Computer malfunctions are difficult to predict as they take place unexpectedly. They cannot be avoided unlike human errors which can be rectified before they occur.
Health risks
Long-term exposure to the technology used in self-driving cars could have health consequences. Autonomous driving technology comprises of remote controls, Bluetooth, Wi-Fi et cetera. These result in a large amount of electromagnetic radiation. For example, complex edge detection algorithms for roads use laser beams which are a form of electromagnetic radiation. Long term and regular exposure to electromagnetic radiation can cause mild to severe health issues such as exhaustion, high blood pressure and migraines.
AI Ethics is a series where we talk about the growing concerns of a lack of regulations and code of conduct for AI enabled products. We aim to present some points and arguments that could be considered for each subject and potentially shape a body of ethical guidelines. Feel free to comment your suggestions or start a discussion below. Hope to see you back with our next post!
Write to us at: the.eclaire.blog@gmail.com
