Central Illinois' Ethics Committee Hosts Driverless Cars Program
On Monday, March 6, 2017, during CPCU Ethics Awareness Month, the Ethics Committee of the Central Illinois Chapter of CPCU presented to a full ball room over lunch at the Chateau Hotel in Bloomington, Illinois, a program on “Driverless Cars: An Interactive Ethical Session.” Kelley Bott, Manager in the State Farm Enterprise Compliance and Ethics Department, introduced the session and Jim Jones, Director of the Katie School of Insurance and Financial Services at Illinois State University, led the group through the ethical scenarios.
Kelley began by emphasizing the CPCU Code of Professional Conduct in ethical considerations. Canon 5 and Rule 5.2 were specifically highlighted by Kelley.
Jim laid out some hypothetical scenarios for autonomous vehicles involving “brake” through technology in programing driverless vehicles. Each table of attendees had a sign indicating which of 4 roles those sitting at the table would play in responding to the scenarios for the new Brakethrough Company, the manufacturer developing the vehicles and their software. Those roles were:
- Engineers whose priorities were the performance of the autonomous vehicles
- Sales/marketing focused on making the vehicles appealing to customers;
- Finance whose main interest was return on investment; and
- Legal/compliance whose priority was avoiding lawsuits, regulatory action and reputation risk.
Jim presented 4 different types of software options which presented different programming and litigation costs. Participants were confronted with ethical dilemmas in selecting these options. While the overall goal was to prevent auto accidents, reducing programming and litigation costs and choosing between options that either increase injuries to pedestrians or passengers create competing interests. Do you design cars to minimize pedestrian injuries or fatalities at the risk of higher passenger injuries or fatalities? Do you design cars destined to be sold in certain states to avoid pedestrian fatalities at the expense of more passenger injuries based on the known higher litigation costs in those states?
The first scenario assumed current braking technology was in use in the company’s driverless vehicles, but that a new swerve program could be introduced to reduce deaths and injuries to pedestrians where the vehicle could not stop in time. Unfortunately, swerving the vehicle would result in more injuries and a few potential deaths in passengers in the vehicles. Jim asked that each of the attendees to respond based on their table’s roles as to whether the new program should be added. A clear majority of the respondents (80%) said yes, led by the legal/compliance and finance groups with over 90% in favor. The engineers and sales/marketing groups were a little more evenly split, but still supportive. There was concern among the sales/marketing groups that passengers are more likely to be household members or family friends that are then more in harm’s way than before making sales/marketing more difficult.
The second scenario added a conditional swerve technology that would swerve only if an empty lane were detected to avoid the pedestrians, but not if the only option were to go onto or over the shoulder. This option would reduce potential injuries and deaths to passengers, but would create a slight uptick in injuries and deaths to pedestrians than would the originally proposed swerve programming. While a simple majority of the respondents supported the new conditional swerve, slightly less than a majority still supported the original swerve programming option. Those supporting only the currently available braking technology under this scenario fell from 20% in the first scenario to only 11% overall. The sales/marketing group expressed the most support for that option due to the concern they expressed earlier, although 70% of that group overall supported a swerve option.
Jim then added to those 3 possible programming options the overall estimated programming and litigation costs as a 3rd scenario. Since the conditional swerve option reduced costs compared to the original swerve mostly due to reduced litigation costs, a 60% majority, up from 50% in the second scenario, favored that option. The engineers and finance groups shifted many of their votes to the conditional swerve option from the original swerve programming. However, the legal/compliance and sales/marketing groups were not swayed by the costs and voted about the same among the 3 options, with legal/compliance preferring the original swerve proposal.
In the final hypothetical, Jim introduced a 4th programming option for a state-modified conditional swerve. That programming has the original swerve option in effect in 2 states with high expected litigation costs for pedestrian fatalities, while the conditional swerve would be in effect in other states. As a result of reducing pedestrian fatalities in 2 states, this program results in lower litigation costs overall. But it consciously increases passenger injuries and potential deaths more in those 2 states than its programming for other states. While this new option received about 25% support overall, its strongest supporter was the finance group while the other 3 groups voted as they had before this option was introduced. No one option at that point had an overall majority supporting it. One table spokespersons said this new option raised the issue of a reputational risk and the morality of deciding whether the technology should injure more pedestrians or passengers based on litigation costs.
Jim concluded by highlighting the Ford Pinto litigation where Ford made decisions based on the costs of recall and remediating the potential for fires in rear-end collisions versus the costs of defending fatalities in fires from those collisions. With our 4 different groups represented in companies like Brakethrough with their different perspectives, but subject to the same Rules of Professional Conduct, who ultimately makes the decision and on what basis?
Published July 5, 2017