News that a self-driving car ‘saw’, but chose to ignore a pedestrian in a fatal collision earlier this year, has thrown the legal questions about this developing technology into sharp contrast.
The self-driving car which killed 49-year-old Elaine Herzberg in Phoenix, Arizona saw the pedestrian as a "false positive", causing its on-board system to decide to ignore her rather than swerve to avoid the crash.
For those who deal predominantly with motoring law – including offences like drink driving, speeding and, in some cases, death by dangerous driving – the development of driverless cars threatens to tear up the rulebook and turn criminal law on its head.
Put simply, where a motoring crime is committed, how much liability can we attribute to a person if the car was carrying out its functions automatically? Can a driver – or perhaps ‘user’ is more appropriate – of a driverless car be guilty of something that the software and mechanics of that device was controlling?
At present, the judicial system here in England and Wales relies upon two basic principles to define an act as ‘criminal’. Firstly, there must be an act or application of force. Second, there must be an intention to carry out that act.
In other words, before a person can be found guilty of an offence, it must be shown that the driver committed the physical act of the crime, and also that they intended to do so.
However, in cases such as speeding, we might expect that the driver suffered a lapse in concentration, or did not see the signage indicating a change in speed limit.
This does not necessarily constitute an intention to act in this way. The law makes allowances for this. In such cases it is not necessary to demonstrate that the individual was willingly, knowingly or even recklessly breaking the law, simply that the act – ie: driving over the speed limit - took place. This is known as ‘strict liability’.
Therefore, should a driving offence be committed by a driverless car – whether intentionally or through carelessness - who would we hold responsible? The driver or user who had no immediate control over the actions of the vehicle? The manufacturers of the vehicle itself?
Supporters of driverless cars have suggested that such concerns might actually be groundless. They point to a world where all manner of sensors, algorithms and satellite navigation mean that vehicles will know where they are, what’s around them and which hazards to plan for, better than any human driver could accomplish.
In short, this would mean an end to motoring crime. A driverless car wouldn’t speed; the intoxication levels of the occupants would be rendered irrelevant.
But as the death of Elaine Herzberg in Arizona suggests, this technology is not infallible. All the sophisticated software in the world cannot accommodate for the incompetent user, the pothole that was unaccounted for or a bug in the system.
There is also the matter of what we do during the transitional phase, when traditional vehicles meet driverless vehicles.
The perspectives of a human driver and a computer programme are likely to be very different, so when incidents happen – as they inevitably will – who would a jury believe? The bus driver or the boffins?
The court cases of the future look likely to be highly complex affairs, with multiple parties involved, each accusing the other of ultimate responsibility.
Yet, it seems unlikely that Google, Uber and other manufacturers would allow themselves to be subject to endless criminal proceedings across endless jurisdictions. The financial and reputational implications would be crippling.
The complex issue of liability is – as of yet – unresolved. Whichever measures are needed in order to address the issues raised by driverless cars, it is clear that the law must adapt. However, in my view there will always be a need to attribute blame in order to maintain a functioning justice system.
Here in the UK, the government – it seems – is taking a similar view. According to a consultation paper by the Department for Transport, any changes in legislation or regulatory framework will be a case of evolution rather than revolution – at least in the short-term.
The government understands that fully-autonomous, door-to-door driverless cars will not be available overnight and we are told that while the driver remains at least partially responsible for the vehicle during the course of its journey, many of the driving offences we are familiar with will remain in place.
The same paper also made a cursory nod to the vast issues surrounding insurance, regulation and the Highway Code, listing proposals to accommodate some of the scenarios mentioned above.
However, given that the paper’s entire section on liability only stretched to three paragraphs the issue is far from closed and the government will be wary of keeping pace with the continued development of autonomous cars before any legal issues arise.
For now, therefore we are only able to speculate. Whether we must wait for answers in further consultations by the government, or if a so-called ‘test-case’ will pass through the courts first, remains to be seen.
By Paul Loughlin, solicitor and motoring law expert at Stephensons