Whisper it quietly, but the machines are coming. Alan Turing developed his eponymous Turing test for Artificial Intelligence (AI) in the 1940s, but it is only now that the technology exists to make some of the devices that early pioneers dreamt of a reality. From driverless cars to the Netflix algorithm suggesting what movies you might like to watch, AI is quietly changing the way we live and the way we do business, and the legal profession is starting to experience some serious changes as technology becomes more affordable and widely available.
University student Joshua Browder has already received many plaudits for his free to use DoNotPay app, which uses a simple algorithm and a few focused questions to help motorists challenge parking fines. All motorists need to do is answer a few simple questions about the nature of the fine, whether markings on the tarmac or appropriate signage was visible, as well as possible points around the manner that the vehicle details were recorded and the penalty notified, and the system will work out some steps that can be taken to challenge the fine. It may seem rudimentary and ridiculously simple to a skilled solicitor to ask these questions about a parking charge, but to the average motorist, particularly in those moments when the red mist has descended after having found the offending penalty notice affixed to the windscreen with a healthy amount of difficult to remove adhesive, it provides a highly accessible and easy to use method of obtaining legal advice to challenge the parking charges. What’s more, it was very popular with average motorists too, particularly as within a matter of months they used it to successfully challenge and overturn over 250,000 penalty notices in London and New York. The cat was out of the bag. Not only was it possible to put such a system together reasonably easily and for a comparatively low cost and minimal resources, it went down a storm with the general public who clearly could not get enough of it. Browder quickly announced plans to roll the service out into other fields where it would not be cost effective to obtain legal advice but where tens and often thousands of people are affected on a daily basis, such as claiming compensation for delayed flights.
The potential is clear. Legal services can be provided, free at the point of use, to the public without the need for expensively consulting a lawyer. The potential size of the market for this ‘un-met legal need’ is immense. It is estimated that nine out of every 10 legal issues and disputes never see a lawyer’s desk or the inside of a court room, with lawyers and courts seen as being expensive, complex and confusing to use. If just a small part of that market can be accessed through the means of AI, the overall size of the legal market could grow exponentially.
However, it would be a grievous mistake to see apps like DoNotPay as the only manifestation of how AI is having an effect in law firms around the world.
The power of Technology Assisted Review (TAR) was powerfully illustrated in Autumn 2016 when the FBI, tasked with the review of thousands of new emails relating to their previously closed investigation into presidential nominee Hillary Clinton’s emails found on a laptop belonging to her assistant, had to review them and make a decision whether or not to proceed with the investigation in a matter of hours under the full glare of the world’s media. In just a few days the FBI confirmed that the emails had been reviewed and that they would not be proceeding with the investigation. This was not achieved through the corralling of a few hundred FBI agents into a warehouse and cancelling the overtime, but rather via the use of technology, or more specifically TAR to allow for the highly accurate review of documents online.
TAR is now an integral part of the e-disclosure toolbox. It might not be appropriate for reviewing all types of evidence. It might struggle to review 3000 hours worth of voicemail messages or 650 text messages full of abbreviations and emojis. But for reviewing documents and emails it can perform the first or second sift of documents easily, learning to discount duplicates and irrelevant material quickly and efficiently, ensuring that the final group of selected documents are passed to the eyes of a human lawyer for review far more quickly - and cheaply - than would otherwise have been the case. The judiciary are catching up to speed as well. Courts now make directions for the TAR of evidence, and will set budgets for the scope and use of such technology within proportionate and sustainable levels.
Law firms are increasingly seeing the value of AI. Dentons has entered into partnership with RAVN Systems to produce an innovative AI application to assist their clients with their questions and concerns over the labyrinthine Brexit process. If this can serve to provide clarity in the sea of confusion over Brexit, with benefits to the wider economy of being able to disseminate clear concise advice to allay the many concerns and fears over Brexit, the system will be able to generate very real and tangible value.
Hold on. Apps giving advice on parking tickets. Machines reviewing documents. Is AI taking work away from paralegals and junior lawyers who would otherwise have been gainfully employed in such roles? It could look like that at first glance until you look a little deeper. A decade ago I recruited and trained teams of paralegals to undertake the review of hundreds of thousands of scanned documents as part of a massive group litigation. The scale of the task was almost unprecedented. Yet, speak to experienced litigators now and you will regularly hear of massive disclosure exercises necessitating the review of huge volumes of data, terabytes upon terabytes of emails, PDF documents, instant messages, notification logs, CCTV footage and voicemail boxes - we aren’t talking about scanned hardcopies anymore. Rather, we are reviewing immense collections of data but still only a tiny fraction of the total volume of data that is now generated every day online in a world of big data that catalogues even the most mundane of tasks for posterity (often in breach of data protection laws but that is a story for another day...). There are now hundreds of actions of this size and scale taking place every year, compared to the few dozen a decade ago, fuelled by big data. So the number of legal jobs has grown, the risk to human lawyers negated by the sheer volume of work.
By satisfying un-met legal need through the use of AI apps, the potential market for those who have a positive experience from using an app is providing a level of public legal education as to rights and legalities that has never previously taken place on such a scale. More people will value the AI advice they receive. They will remember where they received the help and guidance, and if it was received from an AI app from a law firm, they will be back with future legal issues for help, some that an AI app will handle, others that they might be willing and expecting to pay for, with more complex matters requiring a human lawyer. They will have a far more positive and knowledgeable view of the legal profession and the type of help that they need than has been the case in the past. Suddenly the rise of the machines moves from being something to be cautious of to something to welcome with open arms given the potential benefits to the profession, the economy and the general public.
The Law Society worked with the Royal Society to provide evidence and support on their recent project looking at AI in the professions.
Read the full report