26 January 2018
Geofrey Banzi graduated with an upper second class LLB Law degree in July 2017. He is currently working as a Motor Insurance Paralegal for DWF and aspires to specialise as a solicitor operating in the technology and transport industry upon the completion of his LPC. His dissertation was supervised by Dr Catherine Easton.

My dissertation considered the main challenges faced by governments in the process of integrating Autonomous Vehicles (AVs) into their societies, specifically focusing on the Legal, Social, Technological and Regulatory issues arising from AV adoption. The overarching focus was on statutory UK and international law, exploring how the British Government are to enable a seamless transition, from classical paradigms best suited for regulating traditional driving, to one capable of meeting the demands of a driverless society.

AVs are expected to bring many social benefits, including; safer roads, increased mobility for the sick, elderly and disabled, greater utilisation of the strategic road networks (SRN), reduced congestion and reduced carbon emissions. Before these benefits can be crystallised, consolidation and modernisation of existing road law will first be required to bring it in line with the demands of automation. This dissertation specifically focused on the question of legal liability, arising from when products utilising highly sophisticated forms of artificial intelligence (AI), such as AVs, cause harm to humans in the context of road traffic accidents. I made the key contention that current road traffic law under the Road Traffic Act 1988 (RTA) is outdated and in need of clarification.

The legal context focused on the civil law of torts, examining key UK road traffic legislation including the Road Traffic Act 1988, the Vehicle Technology and Aviation Bill as well as the Consumer Protection Act 1987. Key international legislation also covered included the Vienna Convention on Road Traffic 1968, which the UK is a signatory party to but has not yet ratified, as well as its older predecessor the Geneva Convention on Road Traffic 1949. It discussed how the UKs membership to the Geneva Convention rather than the Vienna Convention, which precludes the testing of highly or fully AVs, has enabled the creation of a thriving UK automotive sector in which automakers have been able to gain a crucial first mover competitive advantage under flexible non-binding regulations. These have allowed for the early testing and marketing of highly autonomous capabilities. Yet, whilst acknowledging the benefits of encouraging innovation and AV adoption in the UK, a core argument asserted in this dissertation was that for greater long-term success, the government will need to utilise its crucial policy making role in order to strike the right balance between encouraging innovation whilst also protecting the long term public interest in safety. This will require proactive government action consolidating the current legal framework as well as continued involvement in the development of AVs and their software, ensuring that automakers make the right choices that benefit society as a whole, rather than those purely driven by profit.

Some of the most pertinent questions addressed in the dissertation considered how AVs will in future, at high stages of autonomy, be designed to handle varying road traffic situations where critical life-death decisions will have to be made. In the absence of human instinct, these decisions are expected to be made by ‘perfect’ decision making computer algorithms within the AV. It is not clear whether AVs will be designed to act on the premise of utilitarian decision making i.e the greatest good for the greatest number - which might result in AVs sacrificing their drivers, or conversely, whether they will choose to protect their drivers, at whatever the cost to further human life. This will be an important issue of consideration for policy makers. This dissertation argued in favour of the UK governments increased role in this area, encouraging greater transparency from automakers and possibly even auditing specific AV algorithms to ensure that these design decisions are made appropriately and in the public interest. Other questions also considered ‘who’ or ‘what’ might be classed as the driver of an AV under the current legal framework. If an automated algorithm is in full control of the driving task at hand, such as at stages 4 and 5 of the NHTSA’s 5 stages of automation, how much sense would it still make to designate the title of ‘driver’ to the inactive human operator sat behind the driving seat? Ultimately, by attempting to frame many of the key issues arising from AV use and policy, this dissertation made an overarching argument for much needed legal certainty amidst rapidly changing dynamics and the erosion of the human driver.

Whilst it is easy to argue that because of their expected benefits, which project a utopian-like vision of perfect roads, and ‘drivers’, AVs should be embraced with wide open arms, this dissertation concluded by urging caution in the face of keen optimism, ultimately recognising that whilst being greatly beneficial, AVs and AI will not be completely infallible. Given sheer probability, like all other technological products, they too will be prone to failure. Whilst being largely being in favour of AV adoption, I argued that for the long-term success of its driverless vision, the UK government will need to mobilise itself and act quickly, accounting for the potentially disruptive impact that AV integration will have to current legal, social and economic paradigms. Current Road Traffic Law will need to adapt, to ensure that the UK is best positioned to reap the benefits of automation, whilst also mitigating against future risks.