Claire Wivell Plater (LLB ’81) is managing director of boutique financial services law firm, The Fold Legal. Here she discusses how artificial intelligence may affect the legal profession, with implications for the broader workforce.
Historically, knowledge processing technology has been mainly used for categorising or retrieving relevant material. Useful, but limited. Advances in artificial intelligence (AI) open up a whole new range of possibilities.
Within the legal profession, where I work, the capabilities of this game-changing technology have the potential to significantly reshape the role and functions of many lawyers.
Artificial intelligence is a broad term that includes systems, techniques and technologies, ranging across speech recognition, the processing and translating natural language (natural language processing), learning from examples and precedents (machine learning), sophisticated education of human users (intelligent computer aided instructions) and intelligent problem solving and reasoning (expert systems).
Artificial Intelligence applications do indeed have the potential to undertake some tasks usually done by lawyers, and indeed people in many other professions.
Early attempts to develop artificial intelligence systems largely failed because they assumed the technology would need to model human intelligence, reasoning and general knowledge. As an example, consider this request: ‘When you go to the shop, please bring back a litre of milk, and if they have eggs, get six’. A computer (and probably an engineer) would bring back 6 litres of milk; anyone exercising common sense and general knowledge would bring a litre of milk and half a dozen eggs.
But in 2016, exponential increases in computing power has enabled computers to get the ‘right’ answer using artificial intelligence techniques, rather than trying to mimic human thinking. Computers don’t understand context instinctively in the same way as humans; but by exploiting massive data storage capacity and brute force processing, computers can determine statistically that it’s more likely that ‘six’ relates to the eggs, rather than the milk.
A watershed moment in AI development occurred in March 2016, when a IBM’s DeepMind program, AlphaGo, beat world-class player, Lee Se-dol, in the ancient Chinese strategy board game, Go, which has a game play that offers more possibilities than the total number of atoms in the visible universe. AlphaGo succeeded, not by copying what people do or being taught how to play Go. Instead it ‘learned’ how to play and win by playing millions of games and observing the winning strategies
How is this playing out in the legal sphere?
Codifying human knowledge is generally a starting point for these systems, but they also process and make sense of relatively unstructured data; not unlike what lawyers do. A number of artificial intelligence applications suitable for use in the field of law have already been developed in the UK, US and Australia. Here are some examples:
ROSS is a digital legal expert that helps lawyers power through legal research. Users ask questions in plain English. ROSS then reads through the entire body of law to return a cited answer and topical readings from legislation, case law and secondary sources. ROSS also monitors legal developments and notifies users of relevant changes.
AI software has been developed in the UK to assess the merits of personal injury cases. And in 2014, Michigan State University legal academics created an algorithm to predict the outcomes of U.S. Supreme Court cases. It attained 70 percent accuracy for 7,700 rulings from 1953 to 2013. The technology is now being taken one step further, to predict future litigation.
These technologies are focusing on the big end of town – which is hardly surprising. Only major problems warrant these industry grade solutions. But enterprising companies are also working on smaller scale problems to support individuals and small to medium businesses.
Take Joshua Bowder, an 18 year old Stanford University student from North London. At age 18, he created DoNotPay to help motorists challenge unfair parking fines. Bowder created an AI-based algorithm so users can identify a potential ground of appeal, then the app generates a letter to the appropriate authority.
In the first 9 months, the free app overturned 160,000 parking tickets with a sucess rate of 64%. Joshua’s next project is an app to help Syrian refugees seek legal asylum. It will translate Arabic to English and then draw up the legal paperwork.
It’s understandable that some lawyers may feel threatened by these developments. But if history is any guide, these new technologies are unlikely to completely replace lawyers.
Many of the legal functions being automated are either not currently being undertaken (e.g. monitoring social media posts) or are so expensive that companies are taking the risk of only sampling, rather than doing comprehensive reviews (e.g. due diligence).
So, if anything, these technologies will create greater certainty by reducing risk, and increase access to justice by making some processes faster, cheaper and easier.
And by automating the routine parts of legal work, lawyers will be freed up to focus on more complex, high value areas. After all, what aspiring law student ever imagined they would spend the first few years of their legal career in the e-discovery room collating and highlighting documents? The legal equivalent of a trained chef peeling potatoes!
All that said, more AI is coming and the changes it brings to the whole employment landscape will be profound.
Claire Wivell Plater is a member of the Federal Treasurer’s Digital Advisory Group and is partnering with a local firm to develop an AI application to monitor in real time, multiple social and digital media channels against the legal requirements.