History of Lawyers The first people who could be called lawyers were the great speakers of ancient Greece. Individual people were presumed to present a defense their own cases, but that was circumvented by having a friend better at speaking do it for you. Around the middle of the fourth century, the Greeks got rid of the request for a friend. Second, a more serious obstacle, which the Greek orators never completely overcame, was the rule that no one could take a fee to plead the case of another. This law was disregarded in practice, but was never abolished, which meant that orators could never present themselves as legal professionals or experts. They had to uphold the ruse that they were an ordinary citizen helping out a friend for free, and so they could never organize into a real profession,with professional associations and titles, like their modern lawyers. If one narrows the definition to those men who could practice the legal profession openly and legally, then the first lawyers would have to be the orators of ancient Rome A law enacted in 204 BC barred Roman advocates from taking fees, but the law was widely ignored.The ban on fees was abolished by Emperor Claudius who legalized advocacy as a profession and allowed the Roman advocates to become the first lawyers who could practice openly—but he also imposed a fee ceiling of 10,000 sesterces. This was apparently not much money; the Satires of Juvenal complain that there was no money in working as an advocate.Like their Greek contemporaries, early Roman advocates were trained in rhetoric not law, and the judges before whom they argued were also not law-trained But very early on, unlike Greece, Rome developed a class of specialists who were learned in the law, known as jurisconsults iuris consulti Jurisconsults were wealthy amateurs who dabbled in law as an intellectual hobby; they did not make their primary living from it. They gave legal opinions responsa on legal issues to all comers Roman judges and governors would routinely consult with an advisory panel of jurisconsults before rendering a decision, and advocates and ordinary people also went to jurisconsults for legal opinions. Thus, the Romans were the first to have a class of people who spent their days thinking about legal problems, and this is why their law became so "precise, detailed, and technical.The notaries or tabelliones appeared in the late Roman Empire. Like their modern-day descendants, the civil law notaries, they were responsible for drafting wills, conveyances, and contracts. They were ubiquitous and most villages had one. In Roman times, notaries were widely considered to be inferior to advocates and jurisconsults. ast night I had the pleasure of acting as a discussant, along with Daithi Mac Sithigh, on Mireille Hildebrandt's forthcoming paper Law As Computation in the Era of Artificial Legal Intelligence Speaking Law to the Power of Statistics as well as her Chorley Lecture Paper Law as Information in the Era of Data-Driven Agency. The event was hosted by City Law School as part of its Global Law@City programme. The evening was a great success. The text of my response is below. Response They could be staring us in the face and we just don’t recognise them. The problem is that we’re looking for something very much like us; assuming that they at least have something like the same mathematics and technology…I suspect there could be life and intelligence out there in forms we can’t conceive. Just as a chimpanzee can’t understand quantum theory, it could be there are aspects of reality that are beyond the capacity of our brains. These are the words of Lord Martin Rees, Former President of the Royal Society and current Astronomer Royal following a conference in January 2010 entitled The Detection of Extra-terrestrial Life and the Consequences for Science and Society. Why open my remarks on Professor Hildebrandt’s thought-provoking paper and discussion by reference to alien intelligence and the Astronomer Royal? Well it strikes me that the first thing we must observe is that intelligence may not look like us, may not sound like us and more directly to address Professor Hildebrandt’s thesis, may not identify the same speech acts as we do whether locutionary, illocutionary or perlocutionary. This I believe is the underpinning of the first section of Professor Hildebrandt’s paper. When she talks of Holmes’ famous quote that “the life of the law has not been logic; it has been experience” and from this draws distinction between logic and statistics she is talking of differences not commonalities. All too often when we think of artificial intelligence we think of digitised versions of human brains, us only smarter, quicker, more objective, more precise, more logical. However why should artificial intelligence be human plus, why should it understand our language, our culture or our society? This raises a complex set of questions for us to address as we seek to determine how artificial legal intelligence may interface with human legal experts and practitioners; and with legal rules and norms. To analyse all the implications of this would not be possible within fifteen years much less fifteen minutes so I will restrict my comments to what I see as the central issue raised by Professor Hildebrandt’s paper and discussion: Legal Certainty v Legal Uncertainty and Computation. Legal Certainty v Legal Uncertainty Professor Hildebrandt makes a number of key observations in this area. First that legal certainty “is a matter of statistics rather than logic, of experience and learning rather than framers’ intention or literal textual interpretation. This implies that legal certainty is dynamic and requires hard work to be sustained. If this is done well, the uncertainty that grounds legal certainty is not a drawback but an asset, it makes sure that legal certainty does not trump justice and instrumentality.” To those of us trained in a Common Law background this is apparent. As Lord Mance said in the Oxford Shrieval lecture in 2011: Certainty in the law is highly desirable, but the law must discriminate, sometimes narrowly, between different circumstances. It has in recent years had also, and increasingly, to look to the context and consequences of the language used and decisions reached. As he explains: there is no longer any pretence that the common law has been fixed since time immemorial, or as stated by Lord Reid in The Judge as Law Maker that ‘There was a time when it was thought indecent to suggest that judges make law - they only declare it, but we do not believe in fairy tales any more.’ Thus despite the historic fiction of the “declaratory” view of the common law that it has never changed, and that the judges have merely revealed it from time to time, our modern understanding of Legal Certainty in the Common Law is as understood by Lord Goff in Kleinwort Benson Ltd. v. Mayor etc. of the London Borough of Southwark and Others judicial development of the common law is inevitable, for, otherwise, “the common law would be the same now as it was in the reign of King Henry II”. Professor Hildebrandt observes that “uncertainty is inherent in what Herbert Hart famously described as the open texture of fundamental legal concepts.” She further refers to Neil MacCormick, who points out that “law is an argumentative discipline and no analytic theory of what law is and what distinguishes legal systems from other systems of governance can afford to ignore this aspect of our legal practice, and the distinctive role it plays in a legal system.” This, Hildebrandt observes is an application of speech act theory (which it is): “Law as information must, therefore, be understood as referring to law as a coherent web of speech acts that inform the consequences of our actions, itself informed by the triple tenets of legal certainty, justice and instrumentality that hold together jurisdiction, community and instrumentality.” However it also reflects something else: it reflects the complexity of the legal settlement within a wider framework of regulatory governance. This is a theme touched upon by my colleagues Rob Baldwin and Julia Black frequently over the years, perhaps none more so than in Baldwin’s classic article Why Rules Don't Work. As John Braithwaite explains in his 2002 paper Rules and Principles “by increasing the reliance we can place on a part of the law we reduce the reliability of the law as a whole.” What does Braithwaite mean by this? Well as Rob Baldwin explains rules are drafted and argued about in legal texts without working through how compliance with them is secured in regulatory practice by police, administrative agencies, private litigants and even courts. When flux is great it can be obvious that radically abandoning the precision of rules can allow regulators to increase certainty. To place this into Jurisprudential terms (as Braithwaite does) “Hart points out that rules have a core meaning and a penumbra where their meaning is more uncertain. The more complex and changing the phenomenon being regulated, the wider that penumbra is likely to be; indeed in the most difficult contexts the penumbra of uncertainty swallows up the core creating large numbers of laws that are never enforced.” What does this mean for Hildebrandt’s analysis? She goes on to discuss “Law as computation” and in particular “computation, feedback and performance”. Here two important aspects of Machine Learning are discussed. The first is that Machine Learning is built upon data processing and that data is different from information, which requires meaning. As Hildebrandt says: Neither in the case of organisms nor in the case of intelligent machines does information necessarily imply the attribution of meaning, as it may in the case of humans. Machines work with signs, they do not speak human language; they ‘live’ in a ‘world’ made of hardware and software, data and code. Their perception is limited to machine-readable data; their cognition is based on computation and manipulation of signs, not on generating meaning (though we may attribute meaning to their output and/or their behaviour). Secondly and more importantly for my argument: The idea [of Machine Learning] is simple: to reach its goals the system must reduce uncertainty about the effects of its own behaviour. To reduce such uncertainty it probes its environment and processes the feedback, reconfiguring its processing algorithms until its goal is reached. This leads to the conclusion that “the reliability of Machine Learning depends on a number of inevitable trade-offs, notably between volume, relevance, completeness, accuracy and correctness of the training dataset, the dimensionality and aptitude of the hypotheses space, the time taken for iterant testing, and the availability of the relevant domain expertise.” This places Machine Learning in direct opposition to Braithwaite’s thesis that “As the complexity, flux and the size of regulated economic interests increase, certainty progressively moves from being positively associated with the specificity of the acts mandated by rules to being negatively associated with rule specificity.” As Braithwaite says “When the type of action to be regulated is simple, stable (not changing unpredictably across time) and does not involve huge economic interests, rules regulate with greater certainty than principles.” Braithwaite argues convincingly “most of the things we do in life are simple, stable patterns of action where simple, stable rules can provide us with admirable reliance as to what our legal obligations are.” Thus when we are dealing with simple actions such as the regulation of driving and road traffic or rules relating to qualification for social security, housing or health and safety at work, rules are to be preferred. However once when enter Hart’s penumbra where meaning becomes less certain then for Braithwaite in “complex actions in changing environments where large economic interests are at stake principles are more likely to enable legal certainty than rules.” How is this possible? For Braithwaite there are a number of reasons for this. One is in cases where the regulated activity involves transitional arrangements (such as the regulation of technology) “the regulation of transitional technologies like telecommunications requires “redefinable terms” in rules. What a telephone means today may be something quite different tomorrow.” Another is economic stakes: “by playing the penumbra game, even laws that regulate simple, stable phenomena can be rendered complex.” For myself though the most important factor recalls professor Hildebrandt’s speech act theory analysis. As Braithwaite explains: Flux is a particularly important here. The penumbra of simple rules that regulate stable phenomena is small. It does exist as in Lon Fuller’s example of a statue of a truck in a park where vehicles are prohibited. The unforeseen rare event of the failure to encompass the truck statue in the rule will never cause great problems in regulating the static phenomenon of parks. But in dynamic domains like tax, uncommon things like concrete trucks can be rendered common by gameplaying investors seeking a tax advantage. What does this mean for the application of Machine Learning to legal decision-making? Well while, as Benjamin Alarie has informed us, algorithms get right 98% of the time the answer to the question is a person an employee or an independent contractor? (This is better than the human lawyers who trained the algorithms) This is achieved by reducing the question to a series of algorithmic rules; the removal of uncertainty to reach a certain outcome. Braithwaite though would argue that what is being achieved here is an application of the rules as they are; there is no accounting for flux here. The machine is being asked a static question – is this person an employee or an independent contractor? For Braithwaite this is a type 1 scenario – an action that is simple, stable and does not involve huge economic interests. However the slightest tweak to the stable scenario being tested – a global company such as Uber with in excess of 1 million drivers and an economic imperative to classify them as independent contractors changes the settlement. Now it is a type 2 case one where “complex actions in changing environments where large economic interests are at stake”. Now the economic interests seek to game the system. In response the state enacts more and more rules to plug loopholes opened up by legal entrepreneurs. The thicket of rules we end up with becomes a set of sign-posts that show the legal entrepreneur precisely what they have to steer around to defeat the purposes of the law – or more simply the algorithmic model needed to ensure with 98% certainty that the law can be defeated. Virtual Courtrooms, Virtual Warfare, Virtual Law If one takes a pessimistic view of the future we end up with the worst form of legal settlement: one without Legal Certainty and without the Rule of Law. This is made up the final elements of Hildebrandt’s article: (1) The opacity of Machine Learning; (2) The shift from reason to statistics, and form argumentation to simulation; and (3) judgemental atrophy. If we imagine the Uber scenario operationalized through legal machine learning then we get to the positin where Uber pay a legal consultancy firm millions of dollars to model the perfect scenario where Uber drivers are classified as independent contractors. Meanwhile regulators (including governments) model their own scenarios where the drivers are classified as employees. Millions of dollars on both sides goes into algorithmic modelling and all of the models are protected by IPRs and trade secrecy. We go from argumentation in open court to virtual battles modelled by algorithmic processes hidden from view. If this were to be allowed to occur it would breach at least two of Raz’s principles “All law should be prospective, open and clear” and “The making of laws should be guided by clear rules”. Arguably it may also breach two more “Laws should be relatively stable” and “The principles of natural justice should be observed.” It would also breach Bingham’s first sub-rule “The law must be accessible and, so far as possible, intelligible, clear and predictable”. Finally it seems to be in breach of Lord Chief Justice Hewart’s famous dictum in R v Sussex Justices, ex p McCarthy that “it is of fundamental importance that justice should not only be done, but should manifestly and undoubtedly be seen to be done.” But why would we allow such a position to develop? It is the result of a number of issues raised by Hildebrandt and Braithwaite mixed with a variation of Hildebrandt’s judgemental atrophy argument. There is currently what I would describe as a fetish or perhaps religious fervour surrounding the twin tenets of “big data processing” and “algorithmic regulation or governance”. There seems to be a belief that the marrying of large datasets and algorithmic processing is the answer to almost any question or problem. Luke Dormehl in his book The Formula introduces us to a new definition of self: “the quantified self”. This is the modern self who seeks to quantify decisions: an offshoot of computing power which allows us to crunch numbers very quickly through algorithms. The Quantified Self movement casts its aspirations in bold philosophical terms, promising devotees ‘self knowledge through numbers’. Taking the positivist view of verification and empiricism, and combining this with a liberal dose of technological determinism, the Quantified Self movement begs the existential question of what kind of self can possibly exist that is unable to be number-crunched using the right algorithms?” Yuval Noah Harari develops the religious metaphor: Just as divine authority was legitimised by religious mythologies, and human authority was legitimised by humanist ideologies, so high-tech gurus and Silicon Valley prophets are creating a new universal narrative that legitimises the authority of algorithms and Big Data. This novel creed may be called “Dataism”. In its extreme form, proponents of the Dataist worldview perceive the entire universe as a flow of data, see organisms as little more than biochemical algorithms and believe that humanity’s cosmic vocation is to create an all-encompassing data-processing system — and then merge into it. This movement, the belief that perfect justice can be achieved through algorithmic modelling, will lead to the scenario described. It is probable that law will be reduced to algorithmic profiles in much the same way the Google has reduced library cataloguing of information. Duelling algorithms will take place of duelling advocates but vitally without human experience, language, logic or culture and with a root in rules rather than principles. This leads to the final depressing outcome. As Hildebrandt notes “domain experts may fall prey to gradual deskilling” and argues that it is important that “As law evolves, the software should evolve, and this is where lawyers must come in to make sure fundamental legal principles are taken into account.” However I fear deskilling may not be gradual. I identify three phases of transference of knowledge from human to machine. The first assisted decision-making occurs where digital technology replaces a previous analogue technology or artefact to make our decision-making process simpler. This is a world we clearly already live in. The next phase is supplementary decision-making. This takes place when technology offers us information or processes, which simply were not available before the widespread adoption of digital technologies. This is a world we are now entering. The third phase is autonomous decision-making. This is in the near future where algorithms make decisions without human intervention: this is the world described here. At each of the first two stages we have willingly handed over decision-making autonomy to machines and we will do the same at stage three. I already see students coming through who cannot locate case law except through a search engine or legal database; students who cannot understand the rules of grammar except as enforced by Microsoft Word. Deskilling is happening already. Final words A conflict in values is nearby. The church of Dataism looks to exert control and influence over every aspect of our lives as surely as the Medieval Church did over the lives of its congregations. The Church of Dataism is if anything though much more powerful and influential than the Medieval Church ever was. In Medieval England you could choose to ignore the teachings of the Church in private but Dataism will affect every aspect of our lives from education to transport to law to health care. And the impact of its teachings will be in the very environment itself: to use Lessig’s famous term its code will be law. On the other side sits what I may call the legal luddites. They believe in the rule of law, the value of principles over rules, language, culture, history and logic over statistics. Can both be accommodated in the legal future? Perhaps but the fundamental ideological differences between the two churches give me pause for concern. I thank Professor Hildebrandt for her thought-provoking contribution to the debate.
Demikianlah Artikel Bertemu Langsung dengan Lucinta Luna, Nikita Bocorkan Ciri Khas Transgender
Sekianlah artikel Bertemu Langsung dengan Lucinta Luna, Nikita Bocorkan Ciri Khas Transgender kali ini, mudah-mudahan bisa memberi manfaat untuk anda semua. baiklah, sampai jumpa di postingan artikel lainnya.
Anda sekarang membaca artikel Bertemu Langsung dengan Lucinta Luna, Nikita Bocorkan Ciri Khas Transgender dengan alamat link https://jdsdailyread.blogspot.com/2018/04/bertemu-langsung-dengan-lucinta-luna.html
Postingan
Tidak ada komentar:
Posting Komentar