The Fuse

This Week in AVs: Ford to Test AVs in DC; MIT Study Reveals Culture Attitudes to Trolley Problem; and More

by Alex Adams | @alexjhadams | October 25, 2018

Ford’s AV Testing Comes to Washington DC

This week, Ford announced it will test self-driving vehicles on the streets of The District. In a blog post on Self Driven, Ford’s Medium page for its self-driving endeavors, Ford AV CEO Sherif Marakby said Argo AI has already begun mapping the city’s streets, which is expected to continue into 2019. Marakby said the goal of this trial is to eventually deploy vehicles without safety drivers in DC, adding that Ford will begin building an AV terminal to act as a base in Ward 5. In his blog post, Marakby referenced SAFE’s recent report on AVs, America’s Workforce and the Self-Driving Future, “which found that autonomous technology could improve people’s access to jobs as well as retail markets.”

Moral Responses Captured in AV Dilemma Test

The MIT study revealed that responses varied across cultures, with no global consensus on the preferable option.

The Trolley Problem asks: If an AV is in a situation where a collision cannot be avoided, how should it prioritize where it crashes? The results of a Massachusetts Institute of Technology Media Lab study that crowdsourced this moral dilemma, along with others, were published in Nature this week. Presenting this reimagined version of the trolley problem to millions of people in 10 languages across 233 countries and territories, the study revealed that responses varied across cultures, with no global consensus on the preferable option. As most experts do not believe AVs are capable of making the nuanced distinctions considered in this experiment and it is unlikely that AVs would be forced into such choices, the study’s co-author said he hoped the discussion of ethics moves to risk analysis and about who is at more or less risk, “instead of saying who’s going to die or not,” as well as tackle questions of bias in artificial intelligence.

U.S. DOT Official Calls for Better AV Metrics

The AV industry needs better metrics to show the public and regulators that their products are safe, according to Derek Kan, the undersecretary for policy at U.S. DOT. Speaking at a conference in Washington this week, Kan said the metrics that are most widely used by AV developers—miles driven and the frequency of human intervention—are insufficient on their own to demonstrate the safety of an AV. He noted that not all miles are created equal, with Manhattan a far more challenging prospect than a deserted highway, and human intervention is influenced by driving and engineering choices.

U.S. Takes Bronze in Global AV Preparedness Study

The U.S. is the third best-prepared country for AVs, according to a KPMG study. Achieving a preparedness score of 24.75 out of 30, the U.S. was placed behind Singapore (26.08) and The Netherlands (27.73), with rankings based on four criteria: Policy and legislation, technology and innovation, infrastructure, and consumer acceptance. The U.S. scored highly in the first two categories given its ambitious testing program and the fact that it has the highest number of AV companies worldwide. However, it fell short due to poorer performance in the infrastructure category, as well as the fact that AV technology has poor levels of consumer acceptance in the U.S. To boost consumer confidence in AVs and AV technology, SAFE’s Commission for AV Testing and Safety released a series of best-practice recommendations in January last year, noting the criticality of gaining public acceptance of this nascent technology.