The UK government's announcement last week of the results of its competition for driverless car testing is an exciting development. We expect an announcement of the results of a review of the regulatory regime to follow shortly. As we wait for these results, we identify the top ten areas where we think the law will need to be updated to allow this important technology to move forward.
1 - How to test autonomous critical event control
The tests planned for the UK in 2015 will look at the functioning of driverless vehicles in defined environments and, crucially, with driver back-up for safety reasons.
As we increasingly take the driver out of the picture, vehicles will require systems to determine what action to take in a crisis. (A child running into the road, a car pulling out of a side junction without checking, for example.)
We believe that autonomous critical event control systems require stringent virtual and then ‘controlled environment' testing before any testing on public roads. To test these systems on the roads would inevitably mean members of the public being put at risk. Legislators and policy makers need to determine what the phased testing regime should look like for autonomous critical event control systems to avoid putting the public at unnecessary risk.
2 - How to set the standard for critical event control
With driver error currently blamed for 90% of road traffic deaths, any improvement in safety will be an advantage. But to achieve public trust and acceptance the standards will need to be high. Regulators will need to set safety standards for critical event control systems that are high, but not so high that they are all but impossible to reach.
Autonomous vehicles have the potential to reduce driver error deaths and injuries. So the systems should not have to be perfect; accidents will continue to happen. But the new technology must offer an improved level of safety over the human, error prone, driver.
3 - Rules of the road - how to deal with the requirement for a ‘driver'
Under the old Vienna and Geneva Conventions (the basis of current road traffic laws in many countries), a driver is always needed to be in control of the vehicle. Attempts made so far to soften this requirement to allow for the evolution of autonomous vehicles do not go far enough. An international agreement on the best way forward would be an ideal option, although time-consuming and politically difficult to achieve.
The UK and the US are signatories to the less strict Geneva Convention, and so may be able to satisfy the requirements with the concept of a ‘driver' remote from the vehicle. Lawmakers will need to clarify the question of ‘driver' control alongside introduction of an approved autonomous critical event control system.
4 - Ownership and maintenance
Private ownership seems an outdated concept when we imagine fleets of vehicles moving around cities to serve the needs of everyone (and not only those wealthy enough and appropriately qualified and physically able to drive themselves).
Public ownership or private fleet ownership by manufacturers or specialist corporates is likely to become the norm, and this has a series of consequences. The vehicle owner will need to take over responsibility for its repair and maintenance (organising regular checks is no longer the individual driver's task). And the change in ownership leads on to changes in liability and insurance.
5 - Civil liability – who pays for damage?
For vehicles operating autonomously, the manufacturer is likely to be liable for any accidents caused by defects in the design or functioning of the product. Once a driver is out of the picture, liability for their actions falls away. Only in certain circumstances, deliberate destruction of the hardware, for example, would an occupant of the vehicle be responsible.
Where vehicles are able to operate in both autonomous and non-autonomous modes, the system will need to be carefully calibrated to ensure that the point in time where critical event control changes hands (and therefore responsibility switches from driver to manufacturer) is recorded and determined absolutely. Any vehicle that can operate in an autonomous mode for part of a journey, must have the ability to come to a stop in a safe place if the driver does not retake control when requested to do so by the system.
6 - Criminal liability, and corporate manslaughter?
In the same way, vehicle occupants are unlikely to have criminal responsibility in any but the most extreme circumstances. The current list of criminal offences for anything from speeding to reckless driving would no longer have any place.
But the prospect of criminal responsibility for manufacturers and fleet operators that do not comply with their responsibilities moves up the agenda. Corporate offences and offences on the part of directors would apply to deter wilful flouting of the safety rules.
Legislators will also need to define appropriate criminal liability for interference with both the systems within the vehicles and the systems in the external environment on which the vehicles depend for safe operation.
7 - Insurance
The current system of insurance policies held by individuals would no longer be appropriate. Following on from the increased responsibility on manufacturers and fleet owners would be the need for new insurance products and structures. Manufacturers would need cover for liability arising from defects in the technology, while vehicle owners (fleets rather than individuals?) are likely to need “no fault” insurance to cover any injury or damage caused by any other accidents.
8 - Data protection
As the EU data protection regime, already tough, is toughened up still further (with a new pan-European law expected in 2015) the necessary collection and analysis of personal data involved with ensuring safety, analysing accidents and managing vehicles in public or fleet ownership will have to steer its way through the rules.
9 - Cybersecurity
An integrated set of systems to replace the driver and promote safety would be vulnerable to malicious acts by terrorists or hackers to disrupt its functioning. Regulators would need to take steps to require levels of cybersecurity high enough to counter this threat, and criminal sanctions to deter attack.
10 - The dangers of a slide to autonomy
Perhaps the greatest challenge faced by lawmakers at this important crossroads is to look forward to full autonomy while semi-autonomous and driver-supervised test vehicles take to the roads. As you gradually take control away from the driver (from cruise control to road train systems) you encourage the driver to leave control to the vehicle (and drop off to sleep?). Perhaps the most dangerous prospect is a failure to address the points on the sliding scale between full control and full autonomy without facing up to the safety consequences.
We discuss several of these points in more detail here.