Visiting scholar at Munk School’s Innovation Policy Lab, University of Torono
“There is a way to avoid killing: The car should be programmed to travel at a speed at which such decisions are unnecessary.”
Siraj Ahmed Shaikh
Professor of systems security, Institute for Future Transport and Cities at Coventry University
“Modern automated cars are increasingly complex; the ‘autonomous’ element only serves to add to such complexity. At the core of such a system lies software that connects, controls and makes all the decisions. One contends that the car manufacturer has to take the sole responsibility for any malicious manipulation of such software.
“The sector responds with a simple question: how do we provide security assurance for a system that relies on hundreds of components procured across hundreds of suppliers, and where standards and assurance methods are yet to emerge?
“While the complexity of the engineering process is acknowledged, the party best placed to mitigate security risks is the car manufacturer due to the design and supply chain decisions they take.
Associate producer at The New Yorker
“If billions of machines are all programmed to make the same judgement call, it may be more dangerous to cross the street as, say, an overweight man than as a fit woman. And, if companies decide to tweak the software to prioritise their customers over pedestrians, it may be more dangerous to be beside the road than on it. In a future dominated by driverless cars, moral texture will erode in favour of a rigid ethical framework. Let’s hope we’re on the right side of the algorithm.”
VP marketing, Karamba Security
“Autonomous vehicle developer errors can be automatically mitigated to ensure consumer safety. Autonomous vehicles are promising a great evolution for mobility. This technology requires major upgrades in a vehicle’s internet connectivity and the amount of software needed to make it work. Both of which are potential triggers for developers’ mistakes, ie software bugs. Attackers could exploit those bugs to remotely take control of the vehicle.
“A deterministic cybersecurity technology would prevent the attack and maintain consumers’ safety. Not relying on heuristic analysis allows this real-time decision-making. Implementing cyber security measures during the development cycle provides the built-in embedded security to the vehicle’s controllers.
“The challenge, of course, is how to implement such built-in security without slowing down the AV tight timelines or being prone to security policy mistakes.”
Public transit planning policy consultant, on his website HumanTransit.org
“The fantasy of cars replacing big transit vehicles can lead to serious dystopian outcomes, including higher vehicle miles travelled, higher emissions, and higher exclusion of disadvantaged groups from opportunity. The fantasy is already encouraging neglect of transit systems and opposition to efforts to improve them.”
Michael Sivak and Brandon Schoettle
In Road Safety with Self-Driving Vehicles: General limitations and road sharing with conventional vehicles
“The safety performance of selfdriving vehicles could, in principle, be made perfect. Self-driving vehicles could increase the mobility of those who currently are, for many reasons, prevented from driving. The environmental footprint of self-driving vehicles is envisioned as being much smaller than the footprint of humandriven vehicles because congestion, stop-and-go, and idling could be eliminated or greatly reduced.
“Conclusions: The expectation of zero fatalities with self-driving vehicles is not realistic. It is not a foregone conclusion that a self-driving vehicle would ever perform more safely than an experienced, middle-aged driver. During the transition period when conventional and self-driving vehicles would share the road, safety might actually worsen, at least for the conventional vehicles.
CEO of SpaceX and Tesla, on Twitter
“Yes, excessive automation at Tesla was a mistake. To be precise, my mistake. Humans are underrated.”
Technical manager at ICAEW’s IT faculty
“We need to reimagine public spaces to reduce risks. There is no doubt that automated cars present tremendous opportunities for safer roads. There are also justifiable concerns about safety that will slow down adoption in practice. While we may be holding machines to a far higher standard than humans, demanding close to perfect safety levels, it’s right to challenge the developers to deliver that.
“We should be reimagining public spaces and transport so that they are inherently safer and present far fewer opportunities for collisions.
“Thinking purely in terms of replacing a human driver with an autonomous one, driving on the same roads in the same ways, is missing the point.
“For that reason, I think we are more likely to see autonomous vehicles, perhaps quite soon, in narrow, self-contained environments which are predictable and pretty safe in the first place. Letting them loose on busy city roads, with all the risks encountered there, is a long way off.”
CEO of charity Sustrans, in an interview with Forbes during MOVE expo in London
“Without concerted effort, you end up with an unequal and inequitable distribution of the benefits from the mobility revolution.
“The rich will certainly use AVs first – they are already doing so. But we are going to have to make sure there is dedicated space for people. Dedicated space for people to move under their own power, to walk and to cycle. No matter how clever the technological interface between autonomous vehicles and people, space will need to be dedicated away from motor vehicles. As a civil society we will need to fight for this traffic-free space.
“How do we want to relate to other people? Sitting inside autonomous pods? Or exchanging pleasantries with people as we pass them in the flesh on the street? “There’s little point in technology disruption if it doesn’t make us happier.”
Software engineer, machine learning and artificial intelligence at Google
“I think a lot of the questions being asked in the media about ‘AI’ are really questions about ourselves – about humans and society – that may not be relevant to machines for centuries.”