By Stephen Turner
“We’re not robots Sebastian, we’re physical.” Roy Batty, model number N6MAA10816
The 1982 film Blade Runner, directed by Ridley Scott and starring Harrison Ford and Rutger Hauer, concerns the arrival on Earth in 2019 of a group of renegade Nexus 6 replicants – genetically engineered humans that are superior in strength, agility and intelligence to real humans and yet are so similar in appearance and behaviour that only a long series of questions asked in a cross-referenced sequence can reveal the interviewee as a replicant. The renegade replicants have escaped from an off-world slave colony and murdered their way to Earth in search of a means to neutralise the four-year lifespan that is hardwired into their genetic coding. Harrison Ford plays a blade runner – a future cop – charged with the responsibility of hunting down replicants on earth and ‘retiring’ them.
You see, in the future, replicants are illegal on earth and must be executed on sight!
Whilst the real world technology of 2016 is a considerable way off being able to create the artificial intelligence (AI) of Ridley Scott’s 2019, our legal system will have to adapt to serve the needs of the evolving world in which we live. That world will increasingly include the regulation of autonomous machines and AI – and as AI develops, the involvement of lawyers is inevitable.
Robots and the law: questions for lawyers
The European Parliament has identified the following questions as being currently relevant:
- What is the legal effect of a robot developing autonomous and cognitive features, for example, where a robot becomes an agent and interacts with and shapes an environment?
- What is the legal responsibility arising from a robot’s harmful actions?
- Can a machine be held responsible, wholly or partly, for its actions?
AI, whilst not being human intelligence, must still be subject to regulation and it would be a good idea for lawyers to try to anticipate the legal questions of the future so that we have a body of discussion and debate upon which to draw when looking for the answers. The problem here is that no one has a crystal ball and so anticipating the technology that will come to pass, the applications for which it will be used, and the transformative effect it will have on end users, business, communication and culture is difficult. Who in 2001 predicted Facebook, Twitter and the transformative effect of social media on business, entrepreneurialism, communications and culture?
A more controversial question for the future, one explored in the classics of science fiction is should AI ever be given rights? In Blade Runner, the answer given was ‘no’. The replicants were products of the Tyrell Corporation – used as slaves in dangerous off-world labour roles, in brothels, and in armies during colonisation. However, they are human in all but name – they have emotions, memories, they form bonds, feel desire, love, empathy and grief. As the leader of the replicants, Roy Batty – played by Rutger Hauer, says to one of his genetic designers, “We are not robots Sebastian, we’re physical.”
Whilst the question of AI rights is not one that requires immediate consideration, who is to say that this will not become an issue within the next fifty to one hundred years? Will we one day have robot clients who were not born but were cloned and then grown? Will we see robots as partners? Will we employ them or trust them as our agents? Will they be citizens – with rights and responsibilities? Will we have lawyers who specialise in defending robot rights and championing robot causes?
So many questions, so few answers.
At the present stage of development of robot technology and AI (with AI essentially amounting to intellectual property, software, systems and products) current considerations of partnership extend only to the practical assistance given by AI and the legal issues are broadly those of product liability, including liability in tort and contract. In these respects, lawyers are therefore already turning their attention to the matter of how to regulate the existence and use of AI in the commercial sphere.
The future moves too quickly
A recurring issue for regulation of any new technology is that, due to the time and effort required to take new legislation from first draft to enactment and then to keep it updated, the law often lags behind the pace of technological advancement. As a result, legislation may be superseded by technological advances and outdated provisions may be ill-suited for application to new technology. Lack of fitness for purpose could become a very real problem in the future, as advances in technology are made in rapidly telescoping timeframes. What are we going to do when technology is sprinting forward at 10x and the law is puffing and panting miles behind?
I’m sorry, I don’t understand
Another problem is that legislators may not understand the underlying technology leading to ineffective regulatory measures passing into law. Attempts to regulate infringement of copyright, for example of films and music, whilst not directly related to AI, provide a good example of legislative failure to understand technology. Here is a test for you – let’s see if you understand this…
Copyright is infringed via illegal or unauthorised file sharing, facilitated over the internet via file transfer protocols (software) such as uTorrent. The legislative solution favoured by the UK government is to require internet service providers (ISP), e.g. BT or Virgin, to keep a log of each user’s internet use, including sites visited and the nature of the download or upload traffic. The ISP will then be required to assist the copyright owners in taking enforcement measures, essentially handing over to the owner all data held on the infringer – and the ISP will also act as the initial conduit for communications from the copyright owner. However, copyright infringers invariably use a virtual private network (VPN) which masks both the identity of the user from copyright owner and the destinations that the user visits from the ISP – and a VPN encrypts all traffic up and down the VPN tunnel. The combined effect of this technology is that the ISP will have absolutely no idea where the user goes after they log into their VPN server and the traffic passing from the VPN server, through the ISP and on to the user will be incomprehensible to the ISP.
So be honest, how many of you understood the functioning of the technology that I have just explained? If you did, then you have one up on the legislators. The above example concerns technology that is very simple and easy to understand compared to the vast complexities of robotics and AI. However, legislators in the UK appear oblivious to the ability of VPN technology to thwart attempts to record copyright infringers’ traffic content and their destinations on the Internet. How are we to have confidence in our law makers to legislate at the sharp end of robotics and AI when they will most likely be clueless as to the functioning of that technology?
Tech lawyers – your time is now
If the law is to remain relevant as technology becomes increasingly complex and advances become ever more regular, then as a society we will need lawyers and law makers with an understanding of technology and the imagination, foresight and careful drafting skills to ensure that the law is fit for its future purpose when enacted and for as long as possible thereafter.