Semi-Autonomous cars
Technology can be defined as doing more with less, and with this definition, technology is the basis for the development in prosperity we have experienced in the world. It is easy to look back in time to see what has been a technological advance, but it is far more difficult to predict what will be so in the future. Based on this, it is worth asking the question: "Are (semi-)autonomous cars a good idea?"
A clear financial incentive. To begin, it is interesting to see what drives the development of autonomous cars. ARK Invest, an investment company that focuses on innovation, has become well acquainted with the potential for self-driving cars. In "Autonomous Ridehailing Could Be More Profitable Than We Had Modeled", they explain how the cost of transportation can drop from $3.5 per mile for a regular taxi to $0.25 for an autonomous taxi. This is because the driver can be removed and an estimated 80% reduction in accidents. They put the net present value of this somewhere between 2.1 and 11.9 trillion dollars, depending on assumptions. There is, in other words, a clear financial incentive to achieve this.
The automation paradox. Lisanne Bainbridge writes in "Ironies of Automation" about an automation paradox where the more automated a system becomes, the more important the contribution from those who operate the system also becomes. This happens because it is easiest to automate the simple tasks, and as the system becomes more automated, only the most complex tasks remain. For an operator, this means fewer interactions with the system, but each interaction becomes, relatively, more critical to the system. In relation to autonomous cars, we can talk about a "hand-off", where the driver "takes back" the steering from the autopilot in emergency situations. And these situations require immediate reaction.
Negative effect. Situational awareness can be defined as being aware of what is happening around you and understanding what that information means to you now and in the future. In "Situation Awareness in Future Autonomous Vehicles: Beware of the Unexpected" it is explained how self-driving functionality affects situational awareness of the driver negatively, due to increased confidence and dependence on automation, limited information about the behavior of the system, and reduced cognitive presence by becoming a passive processor of information. An example of this is an emergency situation where the driver first spends time detecting that the functionality/autonomy is not working, then time mentally adjusting to “take back” steering, and then acting. The result is a response that is slower, and worse, to something unforeseen as compared to a driver that had full control all along.
Positive effect. The study "Critical Reasons for Crashes Investigated in the National Motor Vehicle Crash Causation Survey" shows data for accidents for light vehicles from 2005 to 2007 and what led up to the accident. In 94% of the cases, the driver was assigned responsibility for the last incident before the accident. However, the study emphasizes that it is not a fault of the driver, but a lack of recognizing, detecting or performing actions. There is therefore a great need for aids that can assist the driver, and Advanced Driver Assistance Systems (ADAS) is trying to do just that. Some examples are giving warnings before an impending obstacle, calculating braking distance to the car in front, detecting side markings in the road and adjusting the car, detecting pedestrians and cyclists, and monitoring the driver's alertness. This can easily be combined with autonomous functionality to reduce the negative correlation with manual driving skills.
A gradual transition. The main argument for self-driving cars can be summed up in an economic potential, as well as a sharp reduction in traffic accidents. But we also need to get there in the safest, and smoothest, manner.
Erlend