It was a non-obvious thing back in 1996 when the Spice Girls were charting and only 1 in 25 of us were online, to predict the transformation of the legal world through technology, and to do so in depth. Yet this is what Professor Richard Susskind did in his 1996 publication, The Future of Law, much of which is coming now to pass.
Advertisement
What was fascinating last week at Richard’s Society of Computers & Law Annual Lecture, was to hear the very practical path to his realisations at that time, working with Phillip Capper on a system to provide expert guidance on a complicated piece of legislation. The predicted sector transformation may have taken longer than anticipated, but the tell signs were there then, as they are now in way more obvious measure.
And that was the thesis of Richard’s talk: The Future of Law – Reinstated: What if AGI? – his fourth SCL annual lecture in the 15 years he has been its President, and a re-fresh of his 30 year old work. The point being, that while you can’t predict the future only by looking at the past, it is blindingly obvious that something massive is happening and while we get to choose how to respond to that, it would be unwise not to lean in.
Advertisement
The title refers to artificial general intelligence and the talk referenced superintelligence – technology that matches and then surpasses human capability. Where tech capabilities are already beyond what most of us even realise, this question ‘What if AGI?’ is a new call for the legal community to shape or be shaped by what for the technology world is a foregone conclusion – and which some define as already here. Not least, when you look at the scale of investment going into future development and the energy around that.
It seems wild now looking back, that The Law Society wanted Richard to be prevented from speaking in public in the 1990s. Lawyers communicating with their clients via email instead of letter, or undertaking legal research online instead of with books was considered dangerous heresy. Which serves well to demonstrate how deep what Richard calls ‘incumbent resistance’ can run in a profession. We might not see that sort of overt censorship today, but wherever interests are entrenched against innovating in the delivery of client outcomes, then it is the same psychology.
Richard acknowledged the risks of AI – such as models in testing trying to perpetrate blackmail to avoid being switched off – but lingered little upon them, focusing instead on the inevitabilities. These included an increasing trajectory of technology ‘breakthroughs’, reliance on ‘not-yet-invented systems’, and ‘human in the loop’ as an oversight model becoming anathema. His analogy about not having to go to Greenwich to check the time on his watch landed with laughter in the audience.
What did not, I suspect, land so fully, was the message that current narratives have us largely asking the wrong questions. Where we compare ourselves to machines and claim they cannot do what we do – for example exercise judgement, we assume ourselves to be the better solution. Richard presents an alternative inquiry: asking to what problem is human judgement the current solution (i.e. uncertainty). Translated into a legal system context, the question then ceases to become about whether or not a machine can perform like a lawyer or judge, but whether we can build systems that deliver the outcomes clients and society need, fairly, effectively, quickly, affordably, and in a way we can trust. We may nonetheless decide to design in human oversight for its ‘intrinsic value’, even where it is not technically necessary – as an active design choice based on societal sentiment, for example.
Access to justice and the justice system itself has become a central focus for Richard. This is an area in which most agree transformation-level innovation is available and necessary, given the vast volumes, dysfunction and unmet legal need. Where we can apply our collective energy and tech capabilities to that, our lives and jurisdiction will reshape accordingly.
Whatever we do, Richard’s prediction is for ‘fundamental, pervasive, irreversible transformation of our social, political, economic and legal institutions’ precipitated by AGI. His recommendation is to expect materially advanced systems by 2030 – what he calls GenAI+ – and to plan for AGI up to 5 years thereafter.
Alongside a wonderfully mixed crowd, I had the joy of speaking afterwards with a couple of the upcoming generation who attended the talk – both teenagers. Their interest lay particularly in two things. First, the ethical imperative to ensure AI is not used to or for harm. We would do well to heed their views and attention. And second, the influence our use of AI will have not only on work, but on language. The example Richard gave was the Industrial Revolution and how many words and concepts emerged from that time, which themselves shaped society – as will happen with AI, whether having a positive or negative impact. And herein lies the other point: how quickly we get used to things. What feels new or remarkable one month is ordinary the next. Once something is ordinary we stop questioning or thinking beyond it. At its core, the lecture was a call to avoid that, and note that the window to decide what we want from AI and build the systems that will replace deficient ways of working is open now.
So What If AGI? In a profession with so many challenges to resolve and mountains of client demand, the topic can seem too far out to be actionable. Richard’s own path shows otherwise. His predictions came from tackling a practical problem: helping people make sense of one complicated piece of legislation, and in that discovering a universe of possibility. So while we need strategic consideration, collective discourse and more than one dose of dynamic governance on this big topic, the future is actually built right in front of us day by day.
Jenifer Swallow is a strategic legal and technology adviser, former general counsel and former CEO of LawTechUK. She advises legal teams, law firms and technology companies on strategy, innovation and transformation, with a particular focus on AI and the future of legal services.
Legal IT Insider has a two-part podcast with Richard coming up, starting on Thursday 31 September.