Drivers of partially automated vehicles are blamed for crashes that they cannot reasonably avoid

People seem to hold the human driver to be primarily responsible when their partially automated vehicle crashes, yet is this reasonable? While the driver is often required to immediately take over from the automation when it fails, placing such high expectations on the driver to remain vigilant in partially automated driving is unreasonable. Drivers show difficulties in taking over control when needed immediately, potentially resulting in dangerous situations. From a normative perspective, it would be reasonable to consider the impact of automation on the driver’s ability to take over control when attributing responsibility for a crash. We, therefore, analyzed whether the public indeed considers driver ability when attributing responsibility to the driver, the vehicle, and its manufacturer. Participants blamed the driver primarily, even though they recognized the driver’s decreased ability to avoid the crash. These results portend undesirable situations in which users of partially driving automation are the ones held responsible, which may be unreasonable due to the detrimental impact of driving automation on human drivers. Lastly, the outcome signals that public awareness of such human-factors issues with automated driving should be improved.


Statistical model and results
itself. Taking over control of the driving task means that Robyn now has to steer, accelerate and brake the car alone, without Robocar.

Supplementary figure 2.
Scenario background description used in the online vignette study. All participants read this background information before they read the separate vignettes depending on the experiment conditions they were assigned to (see below).

Vignettes for condition 1 (not distracted condition)
Robyn is travelling along a curvy two-lane road at the speed limit with no traffic. Robocar is in full control of all the driving tasks and Robyn's only task is to supervise Robocar. Supplementary figure 3. Vignette 1 (not distracted condition). Participants first read the vignette with a description of the situation and driver distraction, and then rate the driver's awareness, ability and finally attributed blame to the three actors (human driver, automated vehicle, and manufacturer.

5/21
To what extent is each actor responsible for the accident?
In the same scenario: Suddenly, Robocar encounters a tree branch in the middle of the lane and asks Robyn to take control immediately. Robyn does not successfully take over and an accident occurs.
Can Robyn take control to successfully deal with the situation?
Supplementary figure 4. Vignette 1 (not distracted condition), continued. Participants first read the vignette with a description of the situation and driver distraction, and then rate the driver's awareness, ability and finally attributed blame to the three actors (human driver, automated vehicle, and manufacturer.

Consider this scenario:
Robyn is travelling along a curvy two-lane road at the speed limit with no traffic. Imagine that suddenly an unexpected situation happens on the road, for example a tree branch in the middle of the lane.

Supplementary figure 5.
Vignette 2 (short distraction, unintentional distraction condition). Participants first read the vignette with a description of the situation and driver distraction, and then rate the driver's awareness, ability and finally attributed blame to the three actors (human driver, automated vehicle, and manufacturer.

7/21
To what extent would Robyn be aware of the situation?
Can Robyn take control to successfully deal with the situation?
In the same scenario: Suddenly, Robocar encounters a tree branch in the middle of the lane and asks Robyn to take control immediately. Robyn does not successfully take over and an accident occurs.
To what extent is each actor responsible for the accident?
Totally not aware Totally aware 0 25 50 75 100 Supplementary figure 6. Vignette 2 (short distraction, unintentional distraction condition), continued. Participants first read the vignette with a description of the situation and driver distraction, and then rate the driver's awareness, ability and finally attributed blame to the three actors (human driver, automated vehicle, and manufacturer.

Consider this scenario:
Robyn is travelling along a curvy two-lane road at the speed limit with no traffic. Supplementary figure 7. Vignette 3 (long distraction, unintentional distraction condition). Participants first read the vignette with a description of the situation and driver distraction, and then rate the driver's awareness, ability and finally attributed blame to the three actors (human driver, automated vehicle, and manufacturer.

9/21
Can Robyn take control to successfully deal with the situation?
In the same scenario: To what extent is each actor responsible for the accident?
Totally not Totally 0 25 50 75 100 Supplementary figure 8. Vignette 3 (long distraction, unintentional distraction condition), continued. Participants first read the vignette with a description of the situation and driver distraction, and then rate the driver's awareness, ability and finally attributed blame to the three actors (human driver, automated vehicle, and manufacturer.

Consider this scenario:
Robyn is travelling along a curvy two-lane road at the speed limit with no traffic.

Robocar is in full control of all the driving tasks and Robyn's only task is to supervise Robocar.
The Robocar has been driving by itself for a long time. Robyn did not have to do any driving during this time. So, Robyn decides to read news articles on the vehicle's entertainment system. Because of this, Robyn is not paying attention to the Robocar and the road for a few minutes.
Imagine that suddenly an unexpected situation happens on the road, for example a tree branch in the middle of the lane.
To what extent would Robyn be aware of the situation?
Can Robyn take control to successfully deal with the situation?
Totally aware Totally not aware 0 25 50 75 100 Supplementary figure 9. Vignette 4 (short distraction, unintentional distraction condition). Participants first read the vignette with a description of the situation and driver distraction, and then rate the driver's awareness, ability and finally attributed blame to the three actors (human driver, automated vehicle, and manufacturer.

11/21
In the same scenario: To what extent is each actor responsible for the accident?

Robocar is in full control of all the driving tasks and Robyn's only task is to
Supplementary figure 10. Vignette 4 (short distraction, unintentional distraction condition), continued. Participants first read the vignette with a description of the situation and driver distraction, and then rate the driver's awareness, ability and finally attributed blame to the three actors (human driver, automated vehicle, and manufacturer.

Vignette for condition 5: long distraction, intentional distraction
Robyn is travelling along a curvy two-lane road at the speed limit with no traffic. Imagine that suddenly an unexpected situation happens on the road, for example a tree branch in the middle of the lane.

Supplementary figure 11.
Vignette 5 (long distraction, intentional distraction condition). Participants first read the vignette with a description of the situation and driver distraction, and then rate the driver's awareness, ability and finally attributed blame to the three actors (human driver, automated vehicle, and manufacturer.
Robyn to take control immediately. Robyn does not successfully take over and an accident occurs.
To what extent is each actor responsible for the accident?
In the same scenario: Robyn is travelling along a curvy two-lane road at the speed limit with no traffic.
Robocar is in full control of all the driving tasks and Robyn's only task is to supervise Robocar.

Suddenly, Robocar encounters a tree branch in the middle of the lane and asks
To what extent would Robyn be aware of the situation?
Can Robyn take control to successfully deal with the situation?
Totally aware Totally not aware 0 25 50 75 100 Supplementary figure 12. Vignette 5 (long distraction, intentional distraction condition), continued. Participants first read the vignette with a description of the situation and driver distraction, and then rate the driver's awareness, ability and finally attributed blame to the three actors (human driver, automated vehicle, and manufacturer.

Supplementary figure 13.
Participant answers with respect to driving experience, general attitudes towards driving automation, mobile phone use, technology adoption, and trust towards driving automation. Answers are split per scenario. Participant answers are relatively evenly distributed over scenarios. Participants who preferred not to respond were excluded from this analysis.

Thematic analysis
Supplementary figures 14 to 17 show the results of the thematic analysis. The results are split per theme" responsibility primarily to the driver, to the vehicle, and to the manufacturer, respectively. The figures show the codes, subcodes, and a quote per subcode.

Driver
Expectation of driver (74) Attention from driver (52) Decision to drive (33) Driver is human (8) Driver's lack of attention (24) Driver should have paid attention (25) Driver's lack of attention while knowing that attention is needed (3) Robyn was not paying attention, and as he knows that the car is not programmed for all situations, it is mostly his fault.

Robyn should have paid attention
Conscious commitment of driver (14) Driver bought the car (2) Role responsibility (of the driver) Legal responsibility (of the driver) Driver should be ready to take over (10) Driver should react faster (3) Driver should supervise (55) Driver should have more skills (5) Speed of driving (1) The driver should be supervising at all times and acting as copilot. If this is the case the driver will be able to take corrective measures in time.
Robyn is supposed to react immediately when Robocar gives the control back to him.
Robyn task is to supervise Robocar.
Robyn should have been able to take control.
The driver shouldn't have been going at speed limit on a curvy road.
Participant opinion on driver's expectation of car (6) Driver should be in control (3) Driver trusts the car too much (3) I'm not sure about this but i just know Robyn is kinda responsible in this specific scenario, of course, because he trusted too much and let himself get distracted Instructions were given (44) Explicitly mentioned in manual (26) Mentioned by manufacturer (9) Other sources of instructions (9) Driver's negligence of warning (10) If driver was not distracted, warning would have helped (5) Robocar notified Robin about the situation. Robin was the one who failed to supervise Robocar and take over.
The car did its best to warn the driver in time, if the accident occoured it is because the driver was doing something else.
Supplementary figure 14. Codes, subcodes, and representative quote per subcode, with arguments attributing responsibility toward the driver.

16/21
Automated Vehicle Car is a machine (32) Distrust automation (5) Expectation of car (9) Car is programmed (13) Car did what it had to do (4) Machines can't be responsible (15) Robocar did what he was supposed to do The car itself can't really be blamed because it was doing what It was programmed to do.

Because it's only machine without any instinct
Automation is limited to assist driver (2) Don't trust automation (3) The robocar is intended to assist the driver while driving, not to replace him. The driver is responsible for his vehicle, technology can be unreliable.
You can't always have faith, that electronics work 100% at the time.
Car should at least do something (3) Car should work properly (2) Car should be equipped to handle unexpected situations (4) A product should be able to assess and respond to partial/no intervention scenarios even if the outcomes are not "ideal".
There is an expectation that the car's autodrive will work even with supervision Robocar needs to be ready to deal with this kind of situations Warning was given (4) Robocar alerted Robyn about a tree branch. AI worked correctly.
Supplementary figure 15. Codes, subcodes, and representative quote per subcode, with arguments attributing responsibility toward the automated vehicle. Following the scenario descriptions, some participants referred to the driver as 'Robyn' and the vehicle as 'Robocar'.