Vardags — when you need to win

Is it time the legal world embraced AI?

By Thea Dunne -

Is it time the legal world embraced AI?

The legal world has never been known for its speedy uptake of new technologies. While every other industry is anticipating the AI revolution, we’re still filing paper bundles in triplicate, still have judges sitting in postwar magistrates courts, and years away from an online justice system. But with magic circle firms investing more and more in emergent technologies it seems the profession has finally seen the way the wind is blowing.

The culture of Ludditism notwithstanding, the legal industry was always one set to benefit the most from emerging technologies. The administrative burden created by legal proceedings is huge and technology could do much to relieve this, at lower costs to clients. The typing up of attendance notes can easily be supplanted by voice-to-text programs. Other basic tasks too, like collating documents, or pulling together research have been streamlined significantly by developments in computerisation unstructured data organisation. Client interaction too could be disrupted by the advent of virtual-meeting technologies and AI powered virtual assistants and chatbots which are taught to read, reason and communicate.

This repetitive, administrative work signals the way for disruptive technology, but this is just weak AI: machine learning plus big data. We have far more to gain from emerging technologies. The rise of AI powered OCR (Optical Character Recognition) is promising to not only read documents but actually comprehend the contents, with results that suggest it may outperform your typical lawyer. Intelligent software can absorb massive and diverse sets of information to help build persuasive arguments and make well-informed decisions. If it can boost evidence-based decision making in boardrooms or anti-terrorism efforts, why not in the courts?

The old, hackneyed arguments levied against AI - that it’s rigid, rules-based - are being countered one by one. This might have been true with the old symbolic paradigms of the 1980s. AI adhered to set rules - a hard coded, serial symbol-processing system. This allowed for ontological pattern recognition, precise and expressive. One of these early rule-based systems even made a more elegant proof of one of the theorems from Newton’s Principia Mathematica. But it is monotonic, programmed to nail a specific task. Effective with language, the abstract reasoning is there but without context or wider patterns. It doesn’t allow for concept acquisition, has less scope for machine learning. The newer connectionist paradigms changed this. These large networks of simple units are interconnected with links designed to more closely approximate the architecture of the brain. They allow for a different kind of machine ‘thought’: a more teleological structure with the real prospect of concept acquisition. The possibilities for hybrid technologies are fascinating. Between these data-hungry neural networks, weak at high-level symbol processing but with enormous capacity for pattern recognition and machine learning, and traditional symbolic reasoning capabilities, AI may soon have the capacity to both learn and reason.

While the technology is there, there are a number of challenges in building a domain which will encompass legal knowledge. Legal semantics change over time. Epistemological systems are scuppered by the many branches of legal theory which concern developing common law. Legal ontologies, like legal arguments have to contain a mix of legal concepts and real-world concepts. However careful modelling and hybrid technologies could result in a large scale application, if the investment was there.

Writing this, I already hear the protesting cries of lawyers, maintaining that their work could never be automated. Case management demands a diverse skill-set to manage, including at its best empathy, creativity and intuition. Every case throws up its own set of unique demands.

In law, technology may always need to play an assisting role, more R2D2 than HAL 9000. But while we won’t be seeing robo-judges at the bench anytime soon, dismissing new technologies is short-sighted at best and arrogant at worst. The power of smart search, precise logic and big data has already contributed much to the justice system, and AI will only deliver more. Why wouldn’t you harness all the intelligence at your disposal for a case, be it human or artificial. If supercomputers can beat the world chess champion, compose symphonies and assist in diagnosing cancers, I know I’d want them on my side.

If you would like to know more about the issues covered in this article, Vardags offers a free consultation to qualifying individuals.

Our confidential enquiry line is staffed 24 hours, every day of the year. Call 020 7404 9390 today.

Thea Dunne

Thea joined Vardags in September 2016. She read English at Trinity College, Cambridge and now writes as a legal journalist as well as manages Vardags' publicity.