AI in Law

Introduction

It is almost seventy years ago now that Alan Turing wrote the following in an essay – Can Machines Think?

“We may hope that machines will eventually compete with men in all purely intellectual fields. But which are the best ones to start with? Even this is a difficult decision. Many people think that a very abstract activity, like the playing of chess, would be best. It can also be maintained that it is best to provide the machine with the best sense organs that money can buy, and then teach it to understand and speak English. This process could follow the normal teaching of a child. Things would be pointed out and named, etc. Again I do not know what the right answer is, but I think both approaches should be tried .”

Indeed both approaches have been tried. Powerful software products now compete against human professionals and robots are being produced that emulate human emotional responses . There is consensus that there will be a point in time, called the Singularity when machine intelligence will exceed human intelligence; smarter machines will produce even smarter machines and the growth in machine intelligence will be phenomenal. The debate is not whether Singularity will occur but when it will occur. One thing is certain – it will change society in unimaginable ways. The profession of law will not escape.

The environment in which law currently operates and the disputes that it adjudicates has been powerfully impacted by the progress of technology. The incorporation of digital office technology in the everyday running of courts and practices has taken place with active participation of stakeholders. Legal research is easier than it ever was, thanks to the ability to search for key words and phrases. The internet has vanquished the tyranny of distance. Practices are switching over to cloud-based solutions. Researchers have been beavering away for decades, churning together ideas from law, logic and computer science to solve hard problems in law and to automate routine matters. Success for these researchers has been mixed but triumph could be near.

An observer who has been curating the developments in the field would no doubt dispute the imminent arrival of a game-changing technology in the hallowed field of law but there is evidence aplenty that ideas from diverse fields cross boundaries and make their impact felt in incredible ways. The punched card is a good example. At the start of the nineteenth century Joseph Jacquard invented a textiles weaving machine in which the individual warps were lifted in response to location of holes in a punched card and he was able to produce intricately woven patterns more quickly than was hitherto possible. Much later in that century Anselmo Gavioli patented the book organ. He had a number of stiff cards arranged in a fan-fold stack. Music was coded in terms of holes punched in the card. As the card ran at constant speed through the interpreting machine, the hole would let air blow through which in turn made pipes of the organ sound. Charles Babbage took the Jacquard concept to computing, using the cards as data stores and for most part of the twentieth century, punched cards were the dominating way of transmitting commands to a computer. Who would have thought that an idea from weaving will cross into mass produced music and into computing? In a case of reverse takeover computers now electronically control Jacquard machines; computers are an indispensable part of the music production process. Whilst law still adjudicates disputes in the field of intellectual property, technology is close to being able to analyse problems and adjudicate issue.

Technology has crept into the modern life in a big way. Not everyone realises this but technology has been a mere carrier of what is generally labelled as artificial intelligence. The satellite navigation system is not just a repository of maps; it is a specialised intelligent program that based on the vehicle’s current location provides advice as to the best route to reach the desired destination. Word processors not only point out spelling errors but also check for grammatical errors. Chess can be played against a computer and the tax department’s program calculates the tax for a payee based on rules distilled out of the tax legislation. Is artificial intelligence about to make big inroads into law? Yes, if the company Lex Machina is anything to go by. It was cofounded by Prof. Mark Lemley of Stanford University and Josh Becker who has law and MBA degrees from Stanford. The company, now owned by LexiNexis proudly proclaims that it enables lawyers to be better lawyers – much better lawyers. What they did was to bring together patent lawyers and computer scientists to apply the techniques developed for machine learning and natural-language processing to develop a massive database of intellectual property law suits. The data analytic techniques then help deliver a range of answers to its clients, including the probability of success. Although it is a for-profit organisation it provides access to its data for free to courts, government agencies and the academia.

There are formidable obstacles to the widespread use of artificial intelligence in law. For starters, law is meant to be practiced in open courts with public hearings and judgments delivered with clear explanations. An opaque route-mapping system can be tolerated; however, a closed system cannot be permitted in an environment where transparency is extremely important. The common law legal tradition delivers judgment and analysis in a form that is not easily amenable to conversion to structured knowledge that computer scientists aspire towards. There is also the public policy question – who will own the intellectual property in an artificial intelligence application?

On the other hand there are equally powerful factors that are promoting the use of artificial intelligence. Universities and individuals after having success in the medical and engineering fields have taken it as an intellectual challenge to develop systems for application in law. The complexity of laws and legal systems continues to increase so does the cost of complying with the law. For some, the legal process is simply not affordable. The ideal of quick and cheap justice inspires many to find another way.

What impact technology will have on law is not a new question. None less than the Hon Michael Kirby AC CMG wrote in 1998,

“There is no chance that in a quarter century's time, judges either of trial or of appeal, will have been replaced by thinking machines: artificial legal intelligence” .  

He did, however, concede that it was unlikely that the courts will be left completely unaffected by the developments in the field of artificial intelligence. With less than ten years to go, it seems he will be proven right on both counts – but can we be sure?

Practice Management

Obviously, technology can have impact on the practice of law if there are relevant offerings in the market. Already, a number of vendors are offering Practice Management Software (PMS) as a comprehensive tool to manage all functions of a legal practice, namely – case management, billing, accounting, client communications, etc. Typically, the PMS also interfaces with legal databases and search utilities. In 2011, the Queensland Law Society using the services of Kate Hart, a consultant produced a review report covering as many as nine different vendors and interestingly most of them substantially complied with the requirements of a law practice. Of late, PMS has also been available as a cloud-based solution, freeing a practice from the rigours of managing hardware and software infrastructure.

What stands in the way of PMS is paper. Paper records have been a hallmark of modern mercantile society and it would seem that it would only be reluctantly that a change to electronic storage of records would be fully accepted. In fact, the necessity of paper records is enshrined in legislation. Lawyers must keep paper records. For example, trust records must be printed and retained by the law practice. It is probably just a matter of time before it is realised that paper records are more likely to get destroyed than computer records. The willingness of the courts to accept copies rather than originals indicates a shift in thinking. In the notes to s 146 of the Evidence Act 1995 it is written

“Example: It would not be necessary to call evidence to prove that a photocopier normally produced complete copies of documents and that it was working properly when it was used to photocopy a particular document.”

That is, without further a copy produced in the usual manner would be acceptable as evidence and that holds for electronic copies too. It is only a matter of time before the need for physical copies is dispensed with.

There was a time when large commercial organisations had rows of typists’ desks and they could not function without the secretaries buzzing around taking notes in short hand and typing letters with carbon copies. Large engineering offices had likewise rows of drafting stations with many draftsmen poring over them, drawing thick and thin lines and painfully scratching away their mistakes. Secretaries and draftsmen have disappeared, replaced only partially by specialist writers and designers. This change came about due to the increasing use of electronic word processors and computer aided design CAD software. Those who typed their assignments in schools and colleges had no need to dictate notes and were happy to type their own letters. Extrapolation of the trend indicates that the position of legal secretaries will come under pressure.

Professor Gillers, in an article published in 2012 has raised some interesting points in relation to developments in legal practice. He points out that legal practice is now transcending borders. Anyone with access to the internet can now study the laws of another jurisdiction. There is nothing that prevents a lawyer to outsource legal research to a contractor in another country. Lawyers are no longer tied to offices; their offices are mobile and they can practice from anywhere. Self help books and computer programs can assist citizens in do-it-yourself approach to say, write a will. His points are not surprising. The Law Consumer Association in Australia has long been selling DIY Conveyancing and other Kits; going digital will be the logical next step. Accounting companies in USA realised years ago that as long as a USA registered accountant was dealing with the clients and signing off on the tax return, it would not matter who worked on the details. They came up with a winning business plan which has seen a lot of US tax returns being done overseas, for example in Mumbai, India. Lawyers will no doubt chase efficiency and cost reductions and operate across national boundaries. Para-legal personnel: beware.

Technology and Courts

The courts have moved with the times and actively encourage the use of technology. The Federal Court expects the parties to use technology efficiently and effectively in preparation for, and in the conduct of, the trial . The Supreme Court of NSW likewise encourages all parties to consider the prospect of using technology for the purposes of information exchange . The Court requires in the case of electronically stored information like emails, webpages, word processing files, images, sound recordings, videos and databases that discovery and production be given electronically . Electronic discovery or e-discovery has created a problem of its own by requiring that a very large number of documents be gone through even for routine civil matters. Electronic discovery has spawned a new industry - of businesses that purport to reduce the time, cost and scope of discoveries. The challenges posed by e-discovery were anticipated by Baron and Thompson who reacting to the change made to the US Federal Rules of Civil Procedure in 2006, when the definition of document was extended to include electronically stored information, predicted that random sampling and statistical evaluation will play a central role in the process. Their pointers have led to vigorous research and tools from data science are widely used in e-discovery, particularly in the form of supervised learning algorithms targeted at improving quality and reducing cost of review. As mentioned, a new industry is born.

Most courts provide arrangements for hearing by video link. The use of a video link has obvious benefits when one or more of the participants in a hearing (e.g. judge, Full Court, registrar, lawyer or witness) are in different geographic locations . It has been reported that nearly 60 per cent of those who appear in NSW courtrooms appear via video conferencing; meaning there are about 64,000 video court conference sessions in the state per year. The majority of bail matters at the Sydney West Trial Courts are handled on the weekend via video conferencing, and the state has saved tens of millions of dollars over six to seven years.

NSW courts and tribunals conduct increasing amount of work through the NSW Online Registry. The website informs that the court list search service now lets the user search for cases up to three weeks in advance and one week in the past. The searches can also be conducted using apps for Android or iOS devices. A number of forms can be filed online and a new digital service being trialled by the NSW Justice Department is Online Court, which enables legal practitioners and registrars to manage and process preliminary orders without having to enter the courtroom. The service is aimed at saving practitioners and their clients from wasting valuable time at court, waiting for their matter to be heard, as well as the travel costs associated with appearing in person.

Courts being under pressure to reduce waiting times for litigants will increase its reliance on technology to administer justice more efficiently. Online filing is but one example of self service. Self service will increase and administrative staff will be reduced.

Computer Aided Legal Research (CALR)

In 1995, University of Technology, Sydney and University of New South Wales founded the AustLII project and by doing so not only made legal information available to everyone but also made legal research a lot easier. Its databases contain cases and legislation from a number of jurisdictions and it also hosts journals and libraries. Almost 5 million cases have been indexed and can be searched using a number of criteria that include: citation, parties, court, legislation and keywords. In addition searches can also be conducted using Boolean queries. A number of private providers are also active in the field and offer to subscribers powerful search engines that boast of intuitive user interface, filters, drill down methodology and integration of citation with documentation systems.

Legal research has in recent times been also called Computer Aided Legal Research (CALR) owing to the widespread use of computers. CALR remains heavily grounded in keyword search and such does not always deliver well-matched documents and the researcher has to at least cursorily examine every document to see if it is relevant. There is heavy reliance on digital libraries offered on subscription by the likes of Lexis and Westlaw. Stephann Makri found that academic lawyers had poor knowledge of the system and therefore found it difficult to find information. These lawyers were reluctant to attend training classes to improve their research skills despite being aware of the availability of such classes. The reluctance might stem from the fact that vendors of CALR create the impression that CALR is quick and easy and those who struggle with it might mistakenly think that they are inadequate in some way. Nevers makes the case that librarians and not CALR vendors should be teaching legal research, giving it the importance that it deserves.

The training may in future be unnecessary. The shortcomings in digital databases have attracted a lot of interest and development work has targeted the use of data mining techniques. Firdhous at the University of Moratuwa has prototyped a data mining technique to automate legal research . It entails application of a number advanced techniques that facilitate matching of users query with available documents and selected documents are presented in order of the fit. He was able to show that 93% of the time the most relevant document popped up on the top of the list and the results were repeatable, in that 88% of the time the same records were picked up by the search.

Alvin Toffler in his book The Third Wave reminded readers that only speech was natural to humans – reading not, and went on predict that advances in computing power would lead to humans interacting with computer using natural language and not through keyboards and screens. It would be a brave society indeed that would give up on reading and writing but the trend towards natural speech in evident. So called smart phones are experimenting with oral communication and are evidently gaining success. Michael Kirby AC CMG had anticipated this development when he wrote :

The advance of voice recognition continues apace. Within a decade, it seems safe to predict, judges and lawyers will have sophisticated information systems which respond to their voice commands. "Tell me the latest Australian decisions on causation in the law of negligence. Follow this up with recent English and Canadian decisions?" The computer voice will respond at once with the then equivalent to Chappel v Hart. Other case law will roll out whilst the judge or lawyer takes notes or does other things. This is not futurology. It is almost with us. Shops in the United States sell rudimentary voice command toys. Those children, when they become lawyers, will expect nothing less.

Those children do not just expect, they demand that time not be wasted in wading through documents but relevant information brought to them quickly – Google has spoilt every one.

On Line Learning

Online learning is a global phenomenon the impact of which is yet to be fully felt. There is speculation rife that it will completely disrupt the way education is delivered. Coursera , a pioneer in this field offers courses at nil or little cost; there are almost 150 universities including Australia’s best running some 2,000 courses. EdX similarly has almost 1,000 courses on offer and has the likes of Harvard as its founders. It is counter intuitive for respected universities to be offering for free what they can sell at a premium. Realistically though, if they do not have an online presence someone else will step in and offer free courses – there is still a strong sentiment amongst rebels that education and money should not mix. Through their online presence the universities are hedging their bets and if delivery of education abandons the brick and mortar classroom the universities will still be able to operate in virtual classrooms. Amongst the courses on offer are about 75 on legal topics. They are, of course, not fully relevant to Australia but in a globalised world it is of immense benefit to a lawyer to understand the law of other countries. From the learners’ perspective the courses are well prepared, are interesting and whet the appetite for more.

In Australia, there are at least six universities that offer full law courses on line. University of Melbourne Law School and ANU are offering Juris Doctor Degree as a post-graduate course. In May 2011, Central Queensland University launched a course to be taught almost wholly online. The Hon. Michael Kirby AC CMG reviewed the course in an article published in the Australian Bar Review and identified a number of advantages that the course had to offer, namely, the outreach to regional and rural communities, offshore needs and opportunities for ethnic communities and other groups who have not so far been attracted to study law. There is a price to pay. On line education deprives the student of rich campus experience and provides no substitute for student-teacher and peer interactions. These shortcomings are well recognised and Coursera builds an on-line community around the course to partially offset the loss of interaction. For a generation, that spends a lot of waking hours peering into the small screen of a smart phone the new model of interaction might just suffice.

Online resources are not limited to lectures, texts and formal courses. One can, for example read scripts of lectures delivered by Ronald Dworkin or with undiluted joy watch a video recording of his lecture, courtesy the Library of Congress.

The growth in online learning will increase the proclivity of internet-trained lawyers to embrace technology.

Expert Systems

The technology described above could be broadly classified as office products that increase efficiency and enhance quality control. These are not revolutionary products as much the same technology is used in commercial establishments, engineering offices, medical practices, music studios and indeed in almost every walk of life. Of greater interests are advances in technology that enable professionals to automate at least a portion of their work load, leaving them more time to concentrate on more difficult cases. Expert Systems have been around for several decades now and enormous progress has been made in the conceptual and theoretical bases of these systems; however, they are still poorly understood. A layman thinks of these as artificial constructs that will compete against humans. This is a gross misunderstanding. A usual computer program has embedded in it a set of rules that the program repetitively executes and typically a program comprises a set of algorithms and heuristics to solve a class of problems. An Expert System is a computer program of a different kind. It makes a distinction between knowledge and the computing engine that processes this knowledge to provide advice to the human user. This way, the representation and management of knowledge is separated from the coding process . The human experts in the relevant domain pool their knowledge and structure it in a manner that the computational engine can deal with it. The user of the Expert Systems is not being guided by a computer but by the collective knowledge of domain experts.

In May 1997, IBM’s Deep Blue machine beat Garry Kasparov in a chess match to make computing history. Google’s has taken Expert Systems to another level and in January 2016 its AlphaGO defeated Fan Hui, Europe’s reigning champion in the ancient game of Go. Whilst the system still uses about 300 million moves of expert Go players it has been supplemented with machine learning. Cleverly, the researchers behind AlphaGo pitted it against itself and in the process generated a lot more potential moves that were unknown to human experts. The promoters of these projects were not in any frivolous pursuits but were engaged in serious computer science that uses human knowledge as a spring board to go well beyond human capability while also overcoming human fallibility – as an aside, it is said that in one of the man-machine games Kasparov offered a draw as he was scared of committing a blunder - but his opponent harboured no such fears.

In the field of medicine, in the seventies, MYCIN attracted attention. It was not a particularly sophisticated inference engine and used only about 600 rules and worked by asking questions of human operators and using their input as facts to come to conclusions. In about 69% of the cases it was able to identify the bacteria causing severe infections and was able to suggest dosage of antibiotics adapted to patients’ body weights. Remarkably, researchers at Stanford found that it did better than human experts. MYCIN was a technical success and by all indications could have developed into a useful tool to be used in everyday medical practice but it didn’t. The impediment in deploying expert systems, it was found, was the difficulty in extracting knowledge from human experts and converting it into a structured form that can be processed by a machine.

It is sometimes forgotten that the principal activity of business and tax attorneys is not litigation, but rather assisting clients in planning their affairs so as to avoid litigation in the future. Schlobohm and McCarty constructed a heuristic rule-based expert system in 1989 to analyse clients’ needs and recommend a testamentary estate plan. Their system included the relevant tax codes and was goal-seeking and went beyond trouble shooting to creativity.

Game playing and rule based inferring are problems that are well defined and limited in scope and therefore can be solved by computers with high processing speeds. The problems in the field of law are more complex as exemplified by the concept of judicial discretion, particularly in the area of family law. The Family Law Act 1975 is replete with the expression “the court may” which is indicative of discretion; for example, the court may inform itself of a child’s view: s 60CD. The point of allowing discretion to courts, as explained by Deane and Mason JJ in Norbis is that the width of the discretion which Parliament creates maximises the possibility of doing justice in every case. However, they added that the need for consistency in judicial adjudication is an important countervailing consideration in preventing arbitrary and capricious decision-making. The Act puts practical limits on the breadth of discretion through the enunciation of objects and principles: s 60B. It then becomes the duty of the Judge to identify and evaluate competing factors and to give reasons why the particular conclusion was arrived at.

Can computers ever be taught to exercise discretion? Yes, according to Zeleznikow of Victoria University, who along with Kannai and Schild has proposed a model to support decision making in discretionary domains. They suggest that the domain to which a case belongs can be characterised in a number of ways. Where all issues involved in the resolution of the task are known the domain is bounded – for example, property distribution after divorce is done with reference to specified factors, only the weight to be attached to each factor is unknown. On the other hand in refugee law, factors are constantly added to the concept of well-founded fear of prosecution and the problem is unbounded. They define a task to be well defined if all predicates are measurable and their interrelation known – for example, social security law in respect of aged pension. In comparison, property division after divorce is poorly defined. This classification throws a domain into one of four quadrants and a different set of techniques can be applied to each. The bounded and well defined quadrant is amenable to rules based inferring but domains in unbounded and undefined quadrant will call for a mix of technique and yet not deliver satisfactory outcomes. The point of note is this: solutions are close for even hard problems in law.

Artificial Intelligence

Recall that Turing thought of machine intelligence either as a complex problem solver or as a machine imbued with sensors being taught to be human. When the term Artificial Intelligence is mentioned most people envision a human-like machine. Since most robots today fail to see, hear and move as effectively as a human, AI comes a cropper but the ability to sense does not necessarily mean that the machine will be better able to exhibit intelligence. Athletic ability and intellectual ability do not necessarily go hand in hand. If AI is defined to mean a domain-specific problem-solving tool then AI has already surpassed humble humans.

If it is accepted that intelligence means the ability to acquire and apply knowledge and skills, then it is quite evident that humans are born with only a modicum of intelligence. Unless coached and schooled humans can only perform the most rudimentary of problem solving tasks. The skills and tools that we use are artificial constructs. Most languages comprise symbols, phonetic tokens and rules of grammar, which humans have to learn to communicate effectively. Our number system – one of many possible, is an artificial system and humans have to learn the algorithms for addition, subtraction, multiplication, etc. It is not the case that good language and good mathematics comes naturally to humans. To the contrary, humans learn to think in a machine-like fashion and only then are they able to solve problems.

The whole field of law is permeated by artificial constructs. There is nothing natural about say, the concept of trust or insolvency. Even crime is defined circularly – it is what the law terms a crime. Once it is accepted that humans have invented concepts, ideas and algorithms for dealing with societal problems it is easy to come to the conclusion that machine intelligence has a big role to play in debugging the field of law. Law requires a high degree of logical thinking and yet logic is not a core subject in Australian law curriculum; indeed very few universities in the world draw attention to the need to formally study logic. The lack of application of formal rules has led to the situation where legislation is poorly drafted; it is frequently ambiguous, if not self-contradictory as well. If drafted using formal rules of say propositional logic it would be amenable to systematic testing that will reveal all flaws and help eliminate loopholes. Likewise, the difficulty that students and scholars face in discovering the ratio decidendi of a case, would disappear as judges would express themselves concisely in terms of easily understandable propositions. Users well versed in logic would soon realise that they have become more competent at using machines and will begin to enjoy the experience – become better lawyers in the process.

Odd as it may seem, AI and law have a lot in common and this cannot be explained better than this quote from Verheij .

“It is not a coincidence that the fields of AI and Law have crossed paths, as the two fields share method and subject matter. As method, both AI and Law show the powers of what may be called semi-formal modeling. Where the semi-formal models of law take for instance the form of binding precedents and statutory rules, those of AI range from logical representations to robot vehicles visiting Mars. Both AI and Law know that modeling can never be purely formal nor purely informal.  Modeling is always a task of finding the right balance between the order of the formal and the chaos of the informal. In law, rules have exceptions, reasons are weighed, and principles are guiding. In AI, reasoning is uncertain, knowledge is context-dependent, and behavior is adaptive.”

Developing AI for law is a massive intellectual challenge and it is inconceivable that the challenge will remain unanswered.

Drivers of Acceptance

There was a time when televisions were a novelty and people “learnt” to operate them and were interested in how the devices worked. Computers were something that people had to learn and understand but there is already a generation born in the PC era that takes computing technology for granted. They have no fear of the PC just as the generation before had no fear of the TV. Neil Armstrong, the first man to walk on the moon while recounting a conversation with his young son about space travel apparently realised that while he was still looking at moon from the earth, his son born after the moon-landing was looking at earth from the moon. The change in perspective and attitude of succeeding generations has enormous impact on uptake of technology. The PC generation has had its appetite whetted by social media, by satellite navigation, by anticipative advertising and is happy to reside in the digital world. Such of them as enter the legal profession will naturally embrace technology – much the same way as the generation before embraced word processing.

Governments, at least the progressive ones have embarked on the path of openness. Just as AustLII has been hard at work making legislation and case law available to all who wish to access them, the Constitutional Court of Spain has embarked on an e-Justice program. Citizens are able to search a comprehensive database of legal documents and have the same access rights as lawyers and judges. The database is indexed and supplemented with metadata to facilitate research.

Traffic offences are prosecuted automatically with reliance being placed on automation and computer hardware and software. Evidence may be given in proceedings for a speeding offence of a measurement of speed obtained and recorded by an approved traffic enforcement device that is approved for speed measurement . In proceedings for a detectable traffic offence, any one or more photographs that are tendered in evidence, for example for not stopping at a red light are admissible in the proceedings . The machinery underpinning the methodology or automatic prosecution comprises devices with embedded intelligence to detect speed, to detect position, to take a photograph, to transmit data, to match registration plates, to look up ownership records and to apply infringement laws. The point: AI is already embedded and being relied upon and we are barely cognisant of it.

In January 2016, the Commonwealth Government awarded a contract to the value of 2 million dollars to a Dutch company Stichting Neurale Netwerken for the supply of National DNA Investigative Capability. The Australian Criminal Intelligence Commission writes on its website :

“DNA profiling is a forensic identification tool of considerable power. The chance of two unrelated individuals sharing the same DNA profile is approximately one in a billion; making a match between a crime scene profile and that of a suspect a crucial importance. DNA evidence has been used to implicate criminals in serious offences, as well as helping to solve many high volume crimes. Just as importantly, DNA has also clearly established the innocence of many people who might otherwise be suspects for a crime.

These new capabilities can be used to examine DNA profiles to establish biological relationships between individuals, and can be used in the identification of missing persons, unidentified human remains and victims of mass disasters.”

The above contract illustrates the value and reliance that government and societies place on sophisticated technology for prosecution of crimes. In court the DNA evidence is presented by an expert and one can be forgiven for believing that the evidence was solely the result of the expert’s labour. In truth, the expert relies heavily on machine intelligence to identify DNA sequences and for matching them with the records held in the DNA database. Again, society is already relying on machines to determine the guilt or innocence of accused persons.

The need for operating in privacy, which is something that the Internet permits, is prompting people to use web based services. For example, in Holland an online dispute resolution platform was launched in 2015 to help people solve difficult problems like divorce and separation. Research into its efficacy showed that the users, who were engaged in complex conflicts gave a high rating of 7.51 to the website and indicated that they would recommend it to others – growth in usage is likely to be exponential.

Lawyers are witnesses to human follies and foibles and it their constant battle to preferably prevent disputes and failing which to resolve them efficiently. Human intelligence is not always a virtue – much evil has been caused by intelligent beings. Engineers too discovered fairly early in the piece that they could not rely on fellow humans and that is why as much effort goes into making a piece of machinery safe and fool-proof as goes into the functionality of the machine. Turbines have long been fitted with speed governors, trains with dead-man’s handles and boilers with safety valves – all because humans were likely to become inattentive and even negligent. Much as we may appreciate hand-crafted goods, the fact remains that goods produced on automated machinery are better and cheaper. Visual recordings used to be in the form of sketches and paintings but along came photography and it completely displaced the role played by painters and artists; they had to switch to impressionistic and abstract art. The need for quality assurance and risk reduction will continue to push humans back and replace them with infallible machines, even in law.

Social Impact

Richard Feynman, eminent physicist, Nobel laureate, while still a PhD candidate was asked to join Oppenheimer’s team at Los Alamos to develop the atomic bomb. He recounts the events of 6 August 1945 when the bomb was exploded over Hiroshima,

“The only reaction that I remember – perhaps I was blinded by my own reaction – was a very considerable elation and excitement, and there were parties and people got drunk and it would make a tremendously interesting contrast, what was going on in Los Alamos at the same time as what was going on in Hiroshima.  I was involved with this happy thing and also drinking and drunk and playing drums sitting on the hood of-the bonnet of– a Jeep and playing drums with excitement running all over Los Alamos at the same time as people were dying and struggling in Hiroshima.

With regard to moral questions, I do have something I would like to say about it.  The original reason to start the project, which was that the Germans were a danger … But what I did-immorally I would say-was not to remember the reason that I said I was doing it, so that when the reason changes, because Germany was defeated, not the singlest thought came to my mind at all about that, that that meant now that I have to reconsider why I am continuing to do this.  I simply didn’t think, okay?”

There are a lot of people who celebrate and rejoice, pound away at their own set of bongo drums as and when a new breakthrough is achieved and technology melds a little bit more with law. Admittedly, the example of the atomic bomb is rather extreme but it brings into focus some key lessons that Feynman derived. One, to never forget why we are doing what we are doing; two, to do a sanity check and requestion one’s rationale and three, that positive consequences for some are accompanied by disastrous consequences for others. These learnings must not be forgotten.

A prominent evangelist for AI is Professor Richard Susskind who has written several books (he jokingly calls it the same book every few years) on the desirability of using technology in law and the impact that it will have. From his perspective the rationale is clear: it is to provide quick and cheap legal services to the masses and bypass what he calls the gate-keeping function of professionals in contemporary society. He accepts unapologetically that technology will sweep aside professionals and for him the benefit to the masses far outweighs the detriment to the professions, including law. However, that decision, it is submitted is not his to make. It is for the political leaders to make after due considerations.

The impact of technology on society is yet to be fully felt. Technology can simply barge in and sweep aside business models of the past. We have seen it in e-commerce and are seeing it now in the sharing economy where for example, ride-sharing is arguably crowding out the taxi operators. Those, not in denial can visualise what might happen in the field of law. Parties will be able to arrange their affairs – write contracts and wills with the assistance of DIY web-based programs. They will increasingly rely on private law to first establish mutual rights and obligations and second to resolve their conflicts. On the second point, while non-judicial resolution may not be enforceable it will have a strong social backing. Traders on auction sites are more worried about their online rating than profits, because without a good rating there will be no profits. Likewise, those flouting private law will succumb to social pressure and comply; they will not risk being downgraded. Legal offices will outsource – even offshore, legal research; in the quest of efficiency there will be overreliance on machines and the profession will dumb down. Control will pass from legal professionals into the hands of techies.

Society will become dependent on data and electronic communication and individuals will become more vulnerable to stealing of data, manipulation and fabrication of data and those unfortunate to live under suppressive regimes could see their lives shattered by despots who can create false crimes and false evidence to remove obstacles in their way. There would be a multitude of Kafka’s Josef trying to prove their innocence in a digital labyrinth.

As we hurtle towards the Singularity we must pause and think, what kind of society will emerge and what would be the role of professions in that society. Humans are not born with a particular aptitude towards an occupation. During the course of nurturing and learning people discover their competitive strengths and weaknesses and choose occupations that maximise benefits for them. In recent history, people with the ability to learn, memorise and reason have gone into medicine, law and engineering and are reasonably well rewarded. Some argue that they are being rewarded for winning life’s lottery. The very technology that these professions created will wipe out the demand for their professional services. Such people will either be under employed, will compete against robots or veer towards more rewarding professions that we are yet unaware of. What if, in future people stop becoming lawyers and society is bereft of good jurists? Would law still attract the epithet of being a noble profession? That reputation has been earned because there were giants amongst men who stood up against narrow self-interest and ensured that fundamental human rights were upheld. As a case in point, Lord Mansfield said in 1772 :

“The state of slavery is of such a nature that it is incapable of being introduced on any reasons, moral or political; but only positive law, which preserves its force long after the reasons, occasion and time itself from whence it was created is erased from our memory: it is so odious, that nothing can be suffered to support it, but positive law.”

It is on the shoulders of such giants that lawyers and law stand. Technology, mechanisation, automation; whatever name we choose, cannot make the decisions that change the course of history. The independence of judicial reasoning must remain a potent force that at one both resists capricious change and helps humans rise above their animal roots. Society, simply cannot afford to wipe out the pool out of which great judges emerge.

The overriding purpose of legislation and courts’ rules is to facilitate the just, quick and cheap resolution of the real issues in civil proceedings. In contrast, in criminal cases it may be necessary that the prosecution prove the charge beyond reasonable doubt . Sight should never be lost of these fundamental objectives and a risk assessment must be done to ensure that application of technology does not detract from these fundamentals. The government of Western Australia has invested heavily in videolink technology with the ostensible purpose of giving accused persons in remote areas to be heard in a court. The objective is beyond reproach but it remains to be seen whether this measure is simply quick and cheap or will it indeed deliver justice to people who come from an under privileged background and may be more awed by technology itself than the judicial system.

The challenge, quite clearly is for society to assimilate technology but do so in a manner that will minimise social upheaval and it is a formidable challenge.

Embracing Technology

Life on earth has a certain peculiarity. As Dr Robert Winston points out , the two most dangerous chemicals on earth are water and oxygen; water because it is a universal solvent and oxygen because it reacts with almost everything. An alien intelligence form based on, say metal construction will stand no chance on planet earth – it will dare not enter our atmosphere. Peculiarly, life forms on earth have evolved to become totally dependent on these dangerous substances. That is, life has embraced danger and thrives on it. According to Greek mythology, Prometheus stole fire from the gods and gave it to humans. By managing fire humans have distinguished themselves from other animals. We have since tamed the hazardous high pressure steam and nuclear power to improve and extend our lives. Similarly, we will control technology, howsoever intrinsically dangerous it might be.

Society’s reaction to technology has to be measured, restrained and controlled and certainly it should not dance to the tune of bongo drums and it has to be proactive. Bill Gates has counselled , “By embracing the digital age, we can accelerate the positive effects and mitigate the challenges such as privacy and have-vs-have-not. If we sit back and wait for the digital age to come on us on terms defined by others, we won’t be able to do either.” What should then be the terms that should define our engagement with technology? It is submitted that it should be the core human values that are universally cherished that should guide our way. Ronald Dworkin promoted the concept of equality of human beings and equality as a value would probably be accepted without controversy. Equality, according to Dworkin does not mean dividing resources equally amongst people but amounts to the sovereign showing equal concern for all its subjects. Equal concern, in the context of law covers many bases and must be demonstrated in several ways and technology can help – some ideas follow.

Concern for those, who live away from educational institutions and consequently have no access to legal education, mandates that society must promote the dissemination of legal knowledge through non-traditional means. The publication of case law on internet and online legal courses has already made an impact. More could be done, for example by streaming videos of court sessions so that students can learn as to what actually happens in court. Examinations can be held at remote locations under video monitoring to ensure integrity of the examination process.

Concern for those, who cannot access legal advice, requires a number of initiatives. A start has been made by publication of legal information in plain text. There are still people who owing to literacy issues, distance or income are unable to get legal advice. Access can be provided to such people using videolink to lawyers working in community legal centres that are funded by the state. Additionally, interactive programs can be provided such that users can input the particulars of their situation and receive relevant advice.

Concern for litigants in the queue, demands that the objective of quick and cheap justice be promoted using technology. The courts are already on the way to implementing streamlined processes underpinned by technology to efficiently administer justice. Legislation can be made less complex and error free by the use of formal propositional logic and testing the legislation on a software test-bench by applying to a variety of computer-generated scenarios. Litigants can be encouraged to employ a common and standardised research process so that wasteful research and debate is avoided; judges would be asked to adjudicate on a limited number of matters. Witnesses could be asked to give evidence remotely as a standard practice – the modern telephony system permits this already.

Concern for law workers beckons political will. There is no doubt that while technology will create some special positions, it will make many redundant. Susskind’s gatekeepers will need to be redeployed – perhaps more in educating the masses with a view to preventing litigation than in direct adversarial conflict resolution. Things to consider would be multi-professionalism wherein law and another profession are practiced concurrently with law informing the proper course of action in the related discipline. It would pay to bring about a greater match between ethics and law and lawyers of today being replaced by ethicists of tomorrow. Workers could be directed towards reforming and reformulating laws and putting them through test procedures to ensure that the laws are comprehendible and can be applied without problems. In showing equal concern for everyone the society of tomorrow stands a very good chance of redeploying its talented people and will be able to build on its rich legal heritage.

Concern for process: just as laws are periodically reviewed by the Australian Law Reform Commission, so should the use of technology be reviewed. The methodology of how this can be done will itself require careful examination, the key question being – against which standard would computer-assisted justice be evaluated? What weight would be given to speed, cost and just outcomes? It may transpire that people might prefer to use private law and resolution in the future but that does not mean that it renders otiose the state mechanisms of adjudication. The state could and perhaps has a duty to ensure that the private mechanisms are well designed and are fair to all concerned. There should be ongoing risk management of the information technology infrastructure and verification because, pee PG Neumann

“Many of the risks using computers in critical environments stem form people trusting computer systems blindly – not realizing the possibilities that the underlying models are wrong or incomplete. ”

Concern for jurisprudence, necessitates that society nurtures a talent-pool of people well versed in jurisprudence and the laws. This will require that positions are created within universities, governments and courts where these people develop their skills, educate others and help society break out of the rut, when needed and take human society to the next higher level. The development of AI solutions should be carried out by the not-for-profit sector with seed funding provided by the state under oversight of the law societies. This would ensure that AI solutions do not become proprietary and to the contrary are easily available to all users. Researchers and developers would come across inconsistencies and lacunae in laws and would ask the legislature to fix the problems in a proactive manner.

The vision, then is not one of conflict; not one of balancing and mitigating – but one of symbiosis where law and technology serve humanity and serve it together and serve it wonderfully well. If all goes to plan, then -

One day in the fictional future

Prometheus was running late, having spent longer than usual in his daily ablutions but his digital avatar had taken that in its stride and had already ordered the autobot which was waiting patiently when he stepped out of his beachside home. He noted with approval as he stepped into the autobot that Juan, his name for the avatar had already set the climate controls to his liking and the transparency of the windows was just right for the autumn morning. Once Prometheus was safely ensconced and had had the belt tensioned, Juan announced in its androgynous, uni-accented voice, “There is heavy traffic on the roads but the bots have switched to cooperate mode and running bumper to bumper at 70 km/h. The journey to your chambers should take no more than 23 minutes.

There was enough time to get ready for his case. Juan called up Lexomat and within seconds Prometheus was interacting in interview mode, relaying all the facts and issues that his client had conveyed to him. Five minutes later Lexomat reported cheerily, “The key issue is whether the husband is able to claim from the wife the future expenses that might be incurred if the daughter, who is five years old now, when she turns fifteen opts for dental reconstruction and cermet enhancements. There is no binding precedent on this and it would be subject to judicial discretion.” Prometheus quietly contemplated on what had been communicated and multi-tasked watching the news updates.

Coffee was ready when he entered his office. Juan had switched the devices just in time and as he indulged in the pleasure of sipping strong Jamaican coffee he enquired if the opposing counsel was ready for a tele-presence conference. Phoebe had long been ready. She had consulted with Lexomat too and had received much the same advice as Prometheus had. The telepresence connection was quickly established and with mutual control of cameras in each other’s office Phoebe was quick to note that finally her counterpart too had got rid of his formidable legal library and was now firmly in the paperless camp.

It did not take them long to agree that the sensible thing to do would be to ask the court to exercise its decision on the matter. Arguing in an open court had long been passé. They established a telepresence conference with Judge Recht and were surprised to find that she was working from home that day. Earlier in the morning, while bike riding she had run into an autobot and injured her ankle but that was not enough to prevent a workaholic like her from running her court.

“So,” she commences, “both of you agree with the Lexomat analysis and I should admit that I can find no fault with it. On the matter of discretion I would place this case in the same category as fees for private schools. The understanding of the parties prior to the separation and their continuing ability to sustain the expense are persuasive matters in my opinion. On the evidence presented to me it is clear that the parties would have gladly paid for the most expensive orthodontic treatment if they had been together and accordingly I am persuaded that the wife ought not to retract from that position. Please see my associate to have orders framed accordingly. Goodbye.”

Later, while putting on the twelfth hole, Prometheus reflected, we have indeed achieved the objective of quick and cheap justice – finally.