Driverless cars – or ‘automated vehicles’ (AVs), as I was quickly reprimanded by an automotive engineer – are rapidly accelerating into becoming a real commercial proposition. Or are they?
Eager AV enthusiasts are promising that it may not be long before we start seeing these cars as regular additions to our daily commutes. Progress is supposedly fast-paced, with the government already having passed the Automated and Electrical Vehicles Act 2018, extending the existing vehicle insurance framework to cover the use of automated vehicles. But as exciting (or worrying?) as it is to envisage an imminent Blade Runner
future, how close are we to this actually happening? And if it is as close as is being proclaimed, is legislation adapting fast enough to keep pace with the rate at which intelligence is being added to vehicles?
Firstly, a little technical know-how: experts have developed a scale to measure different levels that help to define AVs. This scale ranges from level 0 (no automation, where the driver has complete control), to level 5 (full automation, where no driver is needed). Most current automated cars fall under level 1.However, some manufacturers offer level 2 AVs, where cars can brake and accelerate automatically and, crucially, take control of steering. These ambitious manufacturers pledge that highly automated vehicles will be publicly available within the next few years.
Certainly, the UK seems to be sharing the lead in terms of regular reviews and initial regulations. The government introduced a testing code of practice back in 2015 (the 2015 Code), whereas some EU countries expressly prohibit driverless cars, and many others are yet to approve testing as they try to establish whether such vehicles can be regulated under current laws. As touched on earlier, the UK has already passed legislation under which drivers will now be insured for when they hand over control to their vehicle. So, what about laws for other key areas, such as safety?
Although it is argued that AVs will increase road safety, they do also present several safety questions. In a 2015 review on the subject (setting out the plan to update UK laws to allow for the sale of automated vehicles), the Department for Transport’s focus ranged from smaller matters, such as amending the current MOT procedure so that automated systems are maintained correctly, to addressing real-time high-risk scenarios.
For example, manufacturers create algorithms for their automated vehicles in order for the car to mimic a ‘reasonable’ driver. This includes inputting basic facts so that the car will ‘know’ things such as icy conditions being slippery, or that children may run into the road. Common facts like these can be input easily enough.However, this becomes harder for ethical (but still inherent) concepts, like programming a car to decide how to allocate risk between drivers and, say, pedestrians or cyclists. In a traditional collision, a driver is deemed to have been negligent if he or she did not do what a reasonably competent driver would have done in similar circumstances. How is the law meant to apply this test to AVs? How are manufacturers to create algorithms so cars can mimic the common-sense responses that drivers would have in ethically complicated scenarios, where it is necessary to choose between reducing damage, complying with the law and
allocating risk between those involved in an accident? This is especially pertinent in situations where the law would be trumped by a driver’s common sense, such as mounting the pavement to avoid hitting someone who had run into the road.
In an ongoing review, the government is addressing questions like these, and whether, for example, a higher standard of ‘driving’ should be imposed on an automated vehicle than would be required of a human driver. Currently, there are no mandatory rules regulating this, but only guidelines emphasising the importance of safety. It will be interesting to see whether any future legislation introduced will follow a similar ‘back-seat’ approach, simply imposing a general ‘maximum safety’ requirement, or whether there will be strict regulations on exactly what manufacturers need to include in their algorithms.
There are also possible risks involved in cars which are highly automated (level 3) (in contrast to fully automated (level 5)). Under level 3, the car is able to drive autonomously. However, the driver must be able to take control within a few seconds if needed. Manufacturers consequently view this level as the most dangerous, due to the fact that the driver, although not necessarily driving, needs to remain alert the whole time. The temptation to switch off is all too clear. The 2015 Code covers this, advising that drivers should be trained for the exact moment when the car transitions between automated and manual modes. It also requires that the transitioning system gives the driver sufficient warning of when control becomes necessary. Again, we will have to see how the transitioning mechanism will be regulated when level 3 cars become publicly available. However, it would be surprising if liability issues did not arise – for example, where the driver misses a warning to take back control of the vehicle in time to prevent a collision.
The UK government has eagerly invested in AV infrastructure, and the topic of legality and regulation has been on the agenda for some time. However, many further questions will inevitably need to be addressed and regulated as and when the time comes, such as there being an increased risk to cars from cyber-attacks as they become increasingly dependent on data and electronic communications. Equally, data protection and privacy issues will have to be addressed, with personal data potentially needing to be stored in the event of a collision.
Although not mandatory, the 2015 Code requires current AVs to be able to be controlled by a driver, meaning that the layout of these cars will probably still resemble those in use today (so no Blade Runner
reality just yet). Nevertheless, manufacturers are racing to be the first to produce a near fully automated car, with some hoping to have a level 3 car reach the consumer market in 2021. With all their reviews, consultations and guidelines, let’s hope our lawmakers will be able to keep up the pace, but not lose sight of when to brake!
This article was originally posted in Business Weekly on the 14th March 2019.