Robots & The Law?
Ask most laymen in the street if robots, robotics or AI (bundled together in no small part due to their shared treatment in movies) are subject to any kind of regulation or goals, and a high proportion will probably know of one of the following:
- Do No Harm, aka the Three Laws of Robotics
- Deep Blue, Man vs The (IBM) Machine
- Turing Test, demonstrating human like intelligence
- User-In-Charge, aka UK Automated & Electric Vehicles Act 2018
- Voight-Kampff, from the movie Blade Runner – testing machine Intelligence
The more switched on amongst them may even be aware of initiatives such as the:
- EU Charter on Robotics, aka planned future legislation
- And the ISO (International Organization for Standardization) ISO 10218 standards for Robotic systems
The reality is surprising, the industry has few such regulations in real life and instead relies more on the developers own internal code of ethics to do good and create benefits for mankind….In short, they do tend to adhere to the implied standards mentioned above but aren’t generally compelled to do so legally.
Why is Robotics so unregulated at present?
Aside from the gradually emerging ISO standards and International legislation hoped for in the EU and many countries worldwide, the industry struggles under the burden of aged old testing concepts when it is moving so fast. In part this is because of another well-known concept, Moore’s Law, which observed that the number of processors in a computer chip double every two hence the likely functionality and value improve at similar rates.
Such rapid innovation, which is leading to the advances we are accustomed to in robotics (and its associated AI) is pushing the industry forward but making it increasingly hard to create legislation and standards that will serve us well over time.
So lets take a look at some of these assumed rules impacting the sector.
The Three Laws of Robotics
The Three Laws of Robotics were developed by Russian / USA science fiction author Isaac Asimov, in part to add plot and opportunity to his stories. The rules sought to create an ethical system for humans and robots. The laws first appeared in 1942 and are as follows: “(1) a robot may not injure a human being or, through inaction, allow a human being to come to harm; (2) a robot must obey the orders given it by human beings except where such orders would conflict with the First Law; (3) a robot must protect its own existence as long as such protection does not conflict with the First or Second Law.” Asimov later added another rule, known as the fourth or zeroth law, that superseded the others. It stated that “a robot may not harm humanity, or, by inaction, allow humanity to come to harm.” These rules then are generally considered most relevant for humanoid type autonomous robots and helpers but of course problems occur as soon as the technology branches into areas such as warfare.
IBM’s Deep Blue
IBM’s Deep Blue was the attempt by the computing giant to use AI to outthink a human expert, in this case a chess champion, having already beaten normal players during a program which started in the 1950’s. In 1997, over a contest lasting several days, IBM beat the world chess champion after a six-game match: two wins for IBM, one for the champion and three draws. It was the classic plot line of man vs. machine. Behind the contest, however, was important computer science, pushing forward the ability of computers to handle the kinds of complex calculations such as parallel processing needed to help discover new medical drugs; do the broad financial modelling needed to identify trends and do risk analysis; handle large database searches; and perform massive calculations needed in many fields of science.
The Deep Blue project inspired a more recent grand challenge at IBM: building a computer that could beat the champions at a more complicated game, Jeopardy! and over three nights in 2011, this computer—named Watson—took on two of the all-time most successful human players of the game and was a substantial step forward in computing because it had software that could also process and reason about natural language. Watson demonstrated that a whole new generation of human – machine interactions will be possible.
The Turing Test
The Turing Test was proposed in a paper published in 1950 by British Mathematician, Computing Pioneer, Cryptanalyst, Biologist & Philosopher Alan Turing OBE FRS. It has subsequently become a fundamental motivator in the theory and development of Artificial Intelligence (AI). The test is a deceptively simple method of determining whether a machine can demonstrate human intelligence: If a machine can engage in a conversation with a human without being detected as a machine, it has demonstrated human intelligence.
At the forefront of disruptive technology is the development of Artificial Intelligence and what limitations computers can experience. For this reason, the Turing test was designed to evaluate whether a computer could be “smart” enough to be mistaken for a human. Critics of the Turing Test argue that a computer can be built that has the ability to think, but not to have a mind of its own. They believe that the complexity of the human thought process cannot be coded.
The test is therefore conducted in an interrogation room run by a judge. The test subjects, a person and a computer program, are hidden from view. The judge has a conversation with both parties and attempts to identify which is the human and which is the computer, based on the quality of their conversation. Turing concludes that if the judge can’t tell the difference, the computer has succeeded in demonstrating human intelligence. That is, it can think. More recent development to the test require time limits, multiple judges and a 30% pass rate, but the essence of the test is largely unchanged.
The Turing Test has its detractors, but it remains a measure of the success of artificial intelligence projects. The Loebner Prize is an example of more complex Turing Tests in an annual Turing Test competition that was launched in 1991 by Hugh Loebner, an American inventor and activist.
History of the Turing Test
Several early computers hold early claims to have the ability to have fooled humans in very basic situations. In 1966, Joseph Weizenbaum created ELIZA, a machine that took specific words and transformed the words into full sentences. ELIZA was one of the earliest computers to have fooled human tester into thinking it was human.
Less than a decade later, a chatbot named PARRY was modelled to imitate the behaviour of a paranoid schizophrenic. A group of psychiatrists were asked to analyse conversations with real patients and PARRY conversations. When asked to identify which transcripts were computer programs, the group was only able to identify the machine 48% of the time. Critics of both ELIZA and PARRY state the full rules of the Turing test were not met and do not indicate full machine intelligence.
In 2014, Kevin Warwick of the University of Reading organized a Turing Test competition to mark the 60th anniversary of Alan Turing’s death. A chatbot called Eugene Goostman, who had the persona of a 13-year-old boy, technically passing the Turing Test in that event with a 33% vote from the judges.
In 2018, Google Duplex revealed the capability to performing tasks via the telephone. In various demonstrations, Duplex scheduled a hair appointment as well as called a restaurant, with the human on the other end of the line not realizing they were interacting with a machine. However, critics point out that the interaction does not conform to the actual Turing test and claim the test has still yet to be beaten by a machine.
Recent variations of the Turing Test
Since the creation of the Turing test, more modern approaches have evolved in an attempt to better detect humans and machines. These variations of the Turing test are continually evolving to maintain relevance during technological advancements.
- The Reverse Turing Test aims to have a human trick a computer into having the computer believe it is not interrogating a human.
- The Total Turing Test incorporates perceptual abilities and the person being question’s ability to manipulate objects.
- The Marcus Test has test subjects view media and respond to questions about the content consumed.
- The Lovelace Test 2.0 has test subjects create art and examines their ability to do so.
- The Minimum Intelligent Signal test asks test subjects only binary questions (i.e. only true/false or yes/no answers are allowed).
- User In Charge Legislation represents recent innovative UK law where the legal responsibility for driverless vehicles was considered in the Automated & Electric Vehicles Act 2018, which simply stated that insurers are directly liable for accidents caused by vehicles driving themselves and not passengers being transported in them.
As the prospect of completely driverless vehicles on British roads moves ever close with limited trials already taking place, the Law Commission had been tasked with considering in great detail the issue of legal responsibility for what they term Automated Vehicles. In a far-reaching report (which may indirectly even impact other robotic sectors like manufacturing) published in January 2022, comprising of some 289 pages with 75 new recommendations for fully automated vehicles, they concluded:
Responsibility for Vehicle Manufacturers
In a nutshell, passengers in a vehicle being operated by a fully Automated Driving System (“ADS”), as opposed to a driver support feature such as cruise control, should not be legally responsible for its actions. However, the Law Commission has recommended that under a new Automated Vehicles Act, the manufacturer should be responsible for the actions of automated vehicles and that passengers should be immune from prosecution if the vehicle were to speed, jump a red light, strike a pedestrian or crash. The manufacturer, or Authorised Self-Driving Entity, would be responsible for putting all automated driving features and systems in a vehicle through a two stage approval and authorisation process before the vehicle was permitted on British roads.
Vehicles will need a User-In-Charge
All Automated Vehicles should have a human passenger called a “User-In-Charge” who is qualified to drive and is able to take over in the event of a problem. It is stressed that the User-In-Charge would not need to be actively monitoring the vehicle or what is happening on the road ahead, but would need to be ready to take over after a reasonable time, called the Transition Period, if the Automated Driving System encounters a problem and makes a “Transition Demand” for the human to take over. The User-in-Charge would remain responsible for maintaining and insuring the vehicle, checking all loads are secure before setting off, ensuring all children are wearing seat belts and exchanging details in the event of a collision.
Not all vehicles will need drivers
Some vehicles, such as those used for public transport, may be authorised to drive themselves without a User-In-Charge, and in that situation a licensed “No-User-In-Charge Operator”, in a remote location, would monitor the vehicle. The Operator would need to respond to alerts from the vehicle and ultimately take responsibility if something goes wrong.
The future isn’t here yet
Sadly, the dream of being able to have a few drinks and get your vehicle to take you home in place of a taxi must stay in the realms of fantasy, as the Law Commission states that the User-In-Charge must be fit to drive at all times! Similarly, if you have hopes of sitting back and letting the vehicle take the strain whilst watching your favourite Netflix series or taking a nap then forget it, as the Law Commission has also recommended that the User-In-Charge should not be allowed to use a mobile, screen device or go to sleep!
EU Charter – In various prior studies the EU have already recognised the need for legislation and a comprehensive framework covering robots, robotic systems and AI and these are intended for the future and in 2016 they produced a report looking at the potential need for Analysing the legal value of the Charter on Robotics. However, as mentioned, due to the impacts of things like Moore’s Law, this is an increasingly difficult task and at present they seem hampered by the fact that advances in AI are not yet deemed sufficient for robots make their own judgement call about what constitutes, for example, harm, to humans.
Matters are further complicated with developments such as DNA based nanorobots that are intended to help cure or monitor health issues within humans, or in warfare areas where, say droids or robots are intended to be lethal but will save some lives at the expense of others. These topics all create moral and ethical dilemmas which somehow need to be coded into both robotic systems and legislation.
ISO 10218 standards for Robotic systems — Safety requirements for industrial robots & robotic devices
This is actually real legislation, albeit quiet narrow in definition as follows:
Part 1: Robots.
A robot specific safety specification that addresses manufacturer requirements, functionality, required safety performance, hazards, protective measures and documentation for the robot itself.
Part 2: Robot systems and integration.
Guidance to both end users and robot integrators as it pertains to the safe design, Installation and commissioning of robot systems, as well as recommended procedures, safeguarding and information for use.