Machine learning (ML) systems have very powerful predictive capabilities when given sizeable data sets to work with. Last year, researchers from 1 US and 2 UK universities developed an artificially intelligent method for predicting the outcome of decisions made by the European Court of Human Rights. They achieved an incredible 79% accuracy rating.
A legal team, armed with this type of information, would be in a very influential position when advising clients on whether to settle inside or outside of court, or the likelihood of success down either path, a magic pill for insurance firms deciding whether to defend or pay.
ML based solutions are also strong fits for a number of scenarios which include compliance within an organisation’s communication flows and for comparing large data sets such as contracts to ensure compliance against a standard set, or to simply identify the notable differences which can save a junior lawyer many hours of misery that can be reused more beneficially for both themselves and their firm.
However, contrary to most of the AI news items last year, machine learning is not the only platform capable of providing AI solutions that can aid the legal profession. It is very helpful that the legal profession is embracing and trying to understand the use cases for ML based AI solutions, but other platforms can also intelligently automate their processes.
Cognitive reasoning platforms, which in previous generations were often referred to as expert systems, are also very powerful and can bring a lot of value to a law firm, particularly when they are trying to rescue lawyers from the amount of ‘non-legal’ work they are needing to do.
Cognitive reasoning platforms can generate significant value within a law firm.
Now, some cognitive reasoning platforms use ML techniques to match questions asked against the sizeable data sets fed to them and then come up with what is considered the most appropriate answer. While the response given may have a high degree of confidence, they lack being able to audit to understand why the answer was given and they also lack the ability to interrogate the questioner with qualification questions that will ensure the most appropriate answer is given. They therefore are a tool that can provide tacit legal information but not create knowledge derived from an explicit legal skill set.
Rule based reasoning platforms are able to qualify any questions asked by looking at the knowledge bank available to them, having an understanding of what is required to answer flavours of the question being asked and will guide the questioner through to the specific answer they are seeking.
Particularly important for regulated environments is having an audit trail for the Q&A session and being able to interrogate why specific answers were given. this is not as easily achieved for systems where the computing parameters alone have been coded.
An additional benefit for lawyers is that the rules can be specified in a format that is intuitive and easily understood, reviewed, and even editable by them rather than IT experts. The benefits of this explicit knowledge representation are rapid development and ease of maintenance.