Artificial Intelligence (AI) or cognitive computing simply refers to machine learning technology where computer programs are developed to identify patterns in data for subsequent purposive evaluation. Since specially designed computer programmes are used to understand and replicate human intelligence, AI is hailed as an important breakthrough in the era of Fourth Industrial Revolution. AI application can be 'hard' or 'soft'. In 'hard AI', the machines can theoretically think like humans, whereas in 'soft AI' the machines should be able to do the works that are normally done by human beings. Due to unprecedented projected prospects, technopreneurs around the world have been experimenting the applications of AI in diversified areas. A decade ago, IBM's Deep Blue defeated chess maestro Garry Kasparov. Apple's Siri, Amazon's Alexa, Tesla's automated car, Cogito's phone conversation analysis and real-time guidance, and Google's translation are some of the most powerful AI applications of the present time.
In 1946, an English mathematician, Alan Turing shared the idea of developing intelligence using programming computers. Fifteen years later, California based Patent Lawyer and American Bar Association's Electronic Data Retrieval Committee Chairman, Reed C. Lawlor envisaged that one-day computer would be able to analyse and predict the outcomes of juridical decisions. Legal world powered by AI applications is very close to his prediction in advanced economies as 'soft AI' in the legal sector is widely used there.
In 2016, the researchers applied AI techniques to study 600 judgments of the European Court of Human Rights and revealed that the system could predict the final judgment with 79% accuracy. In the USA, different studies on prediction of court judgments confirmed accuracy between 70-83%, whereas legal experts were accurate by 66% in predicting such cases. Nevertheless, even with such strong accuracy rate, the researchers confirmed that the AI technology, in its present state, can neither replace the human judgment nor be able to work as the substitute of lawyers.
The apprehension that the AI applications will replace the lawyers seems to be a routine negative reaction identical to the previous instances whenever enabling technologies were introduced. For instance, when initiatives were taken to introduce robots in the USA, a census result found in 1950 that around 270 professions would totally be obsolete. However, in the last sixty years, it is found that only one job, i.e. the job of lift operator is completely eliminated, whereas automation has actually increased performance in other jobs in many folds. Thus, the concern that the lawyer will have to sacrifice their works if AI applications are introduced is somehow baseless, rather it has already been demonstrated that it will enable the competent lawyers to work more productively and to do more higher-level and intellectually satisfying task along with the creation of new jobs for the junior lawyers.
Automated decision-making can promote consistency, accuracy, cost-effectiveness and timeliness in the government's decision making. As the application of AI techniques promise to introduce some epoch-making changes, its application is the demand of the time in a country like Bangladesh where hundreds of thousands of cases are pending in different courts of law. A mature AI technology in the legal arena that can analyse the fact and merit of the case and predict the judgment may control the floodgate of bringing trivial matters to the court. Similarly, entry level judges to get an idea of the possible outcome of the cases too.
It is widely circulated that with the vision to become a developed nation, the government has started automation process at various scalein different sectors including law and judiciary. Automation process promises to assist the judiciary work faster and provide speed remedy. Thus, it seems to be a favourable time to actively consider to introduce the AI applications in our legal system so that repeated time and cost can be reduced. Some pilot projects on specific areas, e.g. family, intellectual property, traffic offence or taxation matters can be initiated right now.
But still a long way to go. The policymakers have to face some obvious immediate challenges. For better functioning and performance, relevant data, i.e. various types of legal materials should be digitalised,as the more semi-structured and unstructured data can be made structured, the precise the result can be expected. Language is another serious concern as Bengali is mainly used in the subordinate courts and significant amount of legal materials are still in Bengali. The problem may be solved by using the natural language processing technique. Idea can be taken from the technique developed by Google which has already invented its own language to translate between languages. Regulatory updates and latest court decisions should be incorporated into the system to get better result. Uniform system to deliver judgments may be considered. Most of the government statutory rules and orders are available in pdf format. These should be released in web text format using possibly the same font and size. A concerted and holistic move including all relevant stakeholders is the key behind the success here.
Finally, whatever the immediate reactions of the local law firms or lawyers, the reality is that the future of the legal world like all other sectors will be shaped by the applications of AI even though the technology is still in its infancy. Therefore, the sooner the local legal community will welcome and appreciate this new technology by developing expertise, the better it will be for them. Good to be convinced that the AI techniques are not to replace them but to turn them super charged.
The writer is Senior Lecturer in Law, University of Malaya, Malaysia.