The Supreme Court of India launched its first Artificial Intelligence portal SUPACE (Supreme Court Portal for Assistance in Courts Efficiency) last month. The idea is to leverage machine learning to deal with huge chunks of case data.
During the launch, the then Chief Justice Of India, CJI SA Bobde, said the Supreme Court is embracing Artificial Intelligence in its routine work. He also spoke about how most people gave little thought to AI till the defeat of grandmaster Garry Kasparov at the hands of Deep Blue in 1997.
“It is very interesting that Kasparov won most of the games and lost one. The point is that Deep Blue had not been programmed to understand Kasparov’s moves. The more Deep Blue was used, the more it was employed to play grandmasters, it became better and better and the grandmasters started to lose more games simply because of the phenomenal capacity to analyse without any emotions,” said CJI Bobde.
Sign up for your weekly dose of what's up in emerging technology.
India’s judiciary is mired in backlogs. According to the data available with the National Judicial Data Grid, around 3.81 crore cases are pending in various district and taluka courts in India and more than one lakh cases have been pending for more than 30 years.
SUPACE will be a blend of human and artificial intelligence, and as clarified by Bobde, will not be used in decision-making. The role of AI will be limited to the collection and analysis of data.
“We are not going to let AI spill over to decision-making. The crucial difference being that any knowledge the human being needs to know, whether in fact or law, can be analysed and can be reached to the judge for his decision-making. It fully retains the autonomy and the discretion of the judge in deciding the case, though at a much, much faster pace because of the readiness with which the information is made available by the AI,” the former CJI said.
In 2020, the Supreme Court developed a software called, SCI-Interact, to make all its 17 benches paperless. This software helps judges access files, annexures to petitions and make snotes on computers.
Earlier, the Department of Legal Affairs (DoLA), Ministry Of Law And Justice, has introduced a web-based application called LIMBS or Legal Information Management & Briefing System. The application can monitor cases from high courts and tribunals uploaded by the concerned Commissionerates. The idea is to track the entire life cycle of a case efficiently.
In November 2019, the Apex Court launched an indigenously engineered neural translation tool, SUVAAS, to translate judicial orders and rulings from English to vernacular languages faster and efficiently.
The criminal justice system of the US uses algorithms to estimate the risks of recidivism. Risk assessment tools help judges with information on pre-trial bail, sentencing and parole. Many courts in the country are also actively embracing online dispute resolution (ODR) initiatives.
The most popular and widely used risk assessment tools used in the US justice system are Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) and Public Safety Assessment (PSA).
COMPAS: This tool uses many datasets to predict an offender’s rate of recidivism, risk of violent recidivism, and failure to appear in court. COMPAS breaks it down into static factors including past arrests and dynamic factors including substance abuse, employment history and pessimism. The tool analyses the data to come up with a score. COMPAS does not include any opinions or ambiguous data.
PSA: This tool makes its decisions on nine risk factors such as current arrest, current violent offence, pending charge at the time of the offence, a prior misdemeanour conviction etc. It then weighs each factor and creates a score for individuals. This score predicts the likelihood of a repeat offence.
The Superior Court of Los Angeles’ Gina the Avatar, an online assistant, helps residents handle their traffic citations. The jury bot is built on top of the Microsoft Cognitive Services platform that leverages features like natural language understanding, QnA maker, and translation services. Gina knows five languages and helps over 5,000 citizens a week.
Estonia has introduced an e-residency programme and a national ID smart-card. The Estonian Ministry of Justice has also requested Estonia’s chief data officer to design a robot judge to preside over small claims, disputes of less than 7,000 euros etc.
Unsurprisingly, China has also been adopting AI in the judiciary. The country reportedly has more than 100 robots in courts to recover case histories.
According to author Terence Mauri, in 50 years, machines will recognise physical and psychological signs of deceitfulness with 99.9% precision. “They will be amenable, communicate in all known languages easily and will identify if anyone is lying that couldn’t be recognised by a human,” he said.
For now, in India, SUPACE will be used for administrative purposes and not decision making. “While such attempts in many states shall keep taking place, just like AI can never be an artist, the ability to be fair needs to be evaluated alongside AI technical realities. Automated fairness is not possible to be achieved because ML-based systems do not know how to explain or digest the information they learn, and so a mere idealistic approach to estimate things would not take the initiative further,” wrote Abhivardhan, Chairperson and Managing Trustee of the Indian Society of Artificial Intelligence and Law & the Chief Executive Officer of Internationalism.