The Fuse

This Week In AVs: Partnerships Become Paramount For Industry; AV Transfer Hub Could Cut Trucking Costs; And More

by Anna Polo | July 13, 2020

Autonomous Technology is Becoming a Game of Partnerships
A recent Wired article details how building AVs has become costlier than startups initially predicted, noting that building self-driving vehicles is nearly impossible without a partnership. Karl Iagnemma, CEO of the new joint venture formed between South Korea’s Hyundai and self-driving startup Aptiv, described the unexpected difficulty of building out autonomous vehicles, “We need to integrate them with another very complex system and do it in a way that’s reliable and cost-optimized. It’s really, really hard.”

The list of independent startups is continuing to get smaller.

In addition to Aptiv and Hyundai, partnerships have formed between Waymo and Jaguar; General Motors and Cruise; Argo AI and Ford and Volkswagen; Voyage and Fiat-Chrysler; and Zoox and Amazon. The list of independent startups is continuing to get smaller, as the COVID-19 pandemic intensifies the need for these partnerships with venture capitalists limiting AV investments—making the price of entry into the self-driving industry today around $1 billion. Cofounder of the Israeli self-driving startup Mobileye, Amnon Shashua, shared his opinion to a conference audience in May, “[autonomous driving] is a formidable task, and there are going to be very very few actors who can go from silicon [chips] to self-driving systems.”

Cutting Trucking Costs With Autonomy
As fleet businesses continue to explore autonomous business models to reduce costs and improve performance, new analysis from consultancy Roland Berger estimates one such model known as the transfer hub model could offer per-mile savings of up to 40 percent. In the transfer hub model, conventional trucks are used for the first-mile and last-mile portions while an autonomous truck drives the highway miles—an approach which, according to Roland Berger, will still require more than 1 million drivers to enter the business over the next 10 years.

The transfer hub model could offer per-mile savings of up to 40 percent.

The thought of an autonomous transportation ecosystem consistently raises concerns over job losses, as the role of human drivers becomes less imperative. In response to growing concerns, SAFE examined the economic costs and benefits of AV deployment in a 2018 study. The research concluded there would be a muted impact to the U.S. workforce starting in 2030, returning to full employment by 2050—with total annual benefits by 2050 expected to be $796 billion.

AV Ethics And Alternatives To The Trolley Problem
Requiring people to take the least-worst of two bad options, the “trolley problem” is frequently used to illustrate ethical concerns related to autonomous vehicles and artificial intelligence. Many ethical discussions on AVs focus on whether its technology should be selfish, to protect the vehicle and cargo, or utilitarian, choosing the response that kills the least people. In a recent MIT study, participants were asked to choose between two people an AV should kill, i.e. a driver or a pedestrian, sorting people into different categories and summarizing global moral preferences. However, these studies could have a negative impact on public trust in self-driving cars, as it suggests AVs will be unrealistically influenced according to the ethics of their developers.

An alternative framework has recently been proposed by Veljiko Dubljević of North Carolina State University. In a new paper, Dubljević proposed the Agent-Deed-Consequence (ADC) as an alternative ethical model AVs could use to improve a greater understanding of ethical accountability. The ADC model judges the morality of decisions on three standards: 1) is the agent’s intent good or bad? 2) is the action itself good or bad? 3) is the consequence good or bad? Although self-driving cars are unlikely to come into contact with forced-choice ethical dilemmas, their technology may be developed to include prioritized detection and avoidance of vulnerable road users, such as pedestrians rather than stationary objects.