AI for Lawyers, future shock and an old dog with new tricks

I started working as a lawyer in 1977, prior to my admission and was admitted as a solicitor in 1979. Things have moved a long way from the reel-to-reel tapes on which I used to dictate in the early 1980s.

I had been handpicked for my first job to work on a white collar criminal matter with international ramifications involving what was in those days, Australia’s controlled currency and the alleged breach of Australia’s 1959 Banking (Foreign Exchange) Regulations. I returned to my office from lunch, to see Commonwealth Police surrounding my desk, stuffing my paper files into green garbage bags. I was quickly on the landline, telephoning our Senior Counsel who dictated the terms of an injunction which I wrote out by hand.

I was just twenty-three (23) years of age, in reasonable shape and still lithe and agile, and I ran like Usain Bolt (who wasn’t born yet) to deliver draft orders in Chambers to the Chief Justice of the Equity Division of the NSW Supreme Court. His Honour duly sealed the papers and granted an interim injunction. Then I had to sprint back from the Supreme Court to my office to serve a sealed copy of the handwritten orders on the Commonwealth Police, which stopped them in their tracks.

The following morning there was a hearing in the Supreme Court to determine whether the interim orders should be made final. Not even a fax machine, let alone email, was engaged. Everything was in ink or on paper. Zero in cyberspace.

I recall our first Lanier computer, purchased in 1982 which stored just 256 megabytes. We purchased a printer for almost $26,000 Australian dollars which made so much noise that it required an acoustic hood and a room to itself.

Mobile phones were not smart and they were as big as bricks and then they became very small but were just phones and did nothing else. Briefs to barristers had to be marked in guineas, pounds plus a matching number of shillings (that was just up to 1966). Still after that and for years later, briefs to Counsel were tied with pink ribbon.

Until University, exams were still closed book. Open book exams, calculator – and eventually, computer-access, were progressively introduced. That came as a blow to those of us, endowed with excellent memories, who lost our competitive advantage.

There was much to be said for original thinking, the germination of ideas and initiatives, and the intellectual stimulation borne of spontaneous debate, uninfluenced by unlimited access to information available on the spur of the moment.

With the advent of the internet and word processing technology from the 1980s, agreements became much longer as they became easier to amend.

Lawyers were much readier to compromise when a line change to a document could involve the necessity to retype a whole page, with carbon copies.

The widespread adoption of fax machines only occurred in Australia in the early 1990s and then followed scanners, which were initially only about ninety per cent (90%) accurate. This brought the rest of the world much closer. As a child, I recall sitting in front of the shortwave radio and becoming excited when we could access Radio Moscow or the Voice of America.

The other day, I enquired of an AI platform whether I should use AI technology in reviewing thousands of discovered documents and I was given plenty of reasons why I should but no risks were identified.

Next, I separately questioned my AI service-provider concerning the risks associated with deploying AI to review thousands of discovered documents. Those risks were not evinced until I specifically sought that information.

The risks then listed included data quality issues, bias in algorithms, over-reliance on automation, AI’s struggles with understanding context, particularly faced with complex legal language or documents with ambiguous phrasing, compliance and legal risk regarding data privacy, data security and ethical considerations and security vulnerability, since AI systems can be targets for cyberattack.

The lack of transparency of AI processes makes it difficult to understand how AI makes decisions concerning relevance, which can complicate the review process and make a lawyer uneasy about the thoroughness of AI’s performance and its completeness.

Last week, I made three (3) separate enquiries of two (2) different AI platforms with respect to a question on the requirements of barristers to disclose their fees under our Legal Profession Uniform Law and got three (3) different answers, depending on how I phrased the question.

Only with deeper research could I reconcile the varying responses. I also made an enquiry about the need for commercial lease registration and observed that AI relied on repealed sections of various Acts.

In 1970, Alvin Toffler and his wife, Adelaide Farrell, wrote “Future Shock”, describing the psychological state of individuals and of an entire society, resulting from “too much change in too short a period of time”.

Although the Tofflers predicted much of what would occur over the next fifty (50) years, in fact, technology has developed in post-industrial society, at a more rapid rate than even the Tofflers envisaged.

Most disturbingly, the authors predicted the “death of permanence”, marked by a transient culture where everything ranging from goods to human relationships would be temporary, due to accelerated obsolescence.

Considered along with George Orwell’s “1984”, first published in 1949, in which Orwell examined “the role of truth and facts within societies and the ways in which they can be manipulated”, the Tofflers and Orwell together envisioned a dystopian world which has already become part of our lived experience. Could artificial intelligence have foreseen this?

I asked ChatGPT what the world would be like in fifty (50) years from now and, while enumerating its predictions based upon established learning drawn from the internet, there was no evidence of original or insightful thought.

For a litigator, tactics and strategy are all important. If the plaintiff and the defendant both seek guidance from artificial intelligence, it is likely that they will gain considerable intuition into the approach likely to be taken by their opponent and the matter will proceed in a predictable way, without surprises.

Essentially, litigators can gain an edge on their opponents from the extent to which they can adopt an approach to a problem which differs as far as possible from AI’s prescription.

While we have an adversarial system of law in Common Law countries, judges in deciding cases, will still need to choose between arguable legal positions by reference to precedent.

Already domestic conveyancing and debt recovery, along with corporate filings, are fast becoming superseded fields of legal practice.

If an activity is amenable to a standardised formula, it will not require much, if any, additional human input.

Of course, it will be possible to develop a set of algorithms to ingrain approved values and ethics.

However, when it comes to statute law, criminal justice and sentencing, AI may well be able to be substituted for a judicial officer and algorithms for compassion and leniency, harsh retribution and deterrence, could be devised according to the judicial policy imposed by State legislation, regulation or edict, depending upon the prevailing political ideology in a particular jurisdiction.

Down the track, unpleasant decisions such as whether to hang or not to hang, could be made by artificial intelligence and even the executioner could be made redundant and replaced by a robotic hangman.

There is a range of mental states which humans can access but AI currently cannot: for example conscience, instinct, nostalgia, affection, exhilaration, sexuality, arousal, guilt, shame, cowardice and courage.

If AI leads to mass lay-offs, then unless new industries are created, which require manpower rather than technology, we will be faced with the massive growth of an underclass as a huge burden on the public purse, joining the aged and the infirm, who already absorb an enormous share of the revenue of First World nations.

The expansion of industries based on recreation, entertainment, beauty and pleasure, in which human interaction is uniquely valued, could provide alternative employment for displaced white and blue collar workers.

In Woody Allen’s 1973 sci-fi comedy, “Sleeper”, he introduced a cylindrical booth, designed for immediate technology-induced sexual pleasure, which he called the “Orgasmatron”, invented as a feature of a future society that had largely abandoned traditional human intimacy.

Unless the profits and opportunities produced from the exploitation of artificial intelligence are distributed equitably through the community, those who control AI will enjoy an unprecedented measure of influence and control.

The pursuit of pleasure will become the privilege of the prosperous and the gulf between rich and poor will widen, resembling the imbalance in ancient Athens and Rome, where the common division was between freemen and slaves.

It has never been more important for lawyers to identify the kind of society which our laws should be framed to serve and to enact protections for human rights and the values we consider worthy of preserving.

Whether there is anything that we can do in an attempt to safeguard the universe which, some of us had come to cherish, is an open question that only Artificial Intelligence is likely to be entrusted to resolve.