By Aman Kumar Yadav, fourth year student (5 year B.A.LL.B.), National Law Institute University, Bhopal
Introduction
Robots, Artificial Intelligence (AI), drone are not buzzwords anymore, courtesy of COVID-19. The pandemic has affected many lives and livelihoods across the globe, and various organizations whether small or large, are finding new ways to operate effectively and survive the current crisis. Machine learning technology is now playing an important role in various fields be it healthcare, contact tracing & data dashboard, e-commerce, etc. That being said, let’s examine the scope of this technology in Indian judiciary.
To mark the celebration of the National Constitutional Day (November 26), 2019, the Chief Justice of India, Justice Bobde during his speech announced the launch of a neural machine translation (NMT) tool called Supreme Court Vidhik Anuvaad Software (SUVAS) that can translate orders and judgments in English to nine vernacular languages. Furthermore, he expressed his optimism towards the use of AI to lessen its perennial backlog of cases and improve the overall efficiency of the Indian judiciary.
One of the major advantages of AI lies in its efficiency. As of November 2019, there were around 59,867 cases pending in the Supreme Court, 44.75 lakh cases in various high courts and 3.14 crore at the district and subordinate court levels. The more shocking fact is that these numbers don’t fluctuate much, as the pendency in December 2014 stood at similar numbers in these respective courts. [1] AI could help in clearing this backlog, especially in cases of a repetitive nature and in document management. It can also help in varied legal research and due diligence work with significant improvement in productivity and accuracy.
While the developments taking place on the technological front, there is no comprehensive legislation to regulate AI or machine learning in India yet. However, some signs show the government’s focus on the technological advancements in the field of machine learning and AI as is evident from Union Budgets of the years 2019 [2] and 2020 [3]. Also, in June 2018, NITI Aayog released a policy paper, ‘National Strategy for Artificial Intelligence’, which considered the importance of AI in different sectors. In July 2019, Ministry of Electronics and Information Technology (MeitY) released extensive reports of its four committees [4] constituted to work on a policy framework to promote and regulate AI in India.
That being said, there are two major questions that we face, (a) whether there are any legal challenges related to AI replacing human judges, and (b) whether we are ‘technically’ equipped with sufficient data to build an effective AI-powered judicial system.
Legal Challenges
It can be argued that the AI wouldn’t be biased per se as it is based on the predictive algorithm and can help in making law more uniform as it removes uncertainty brought by personal bias of different judges. However, it’s not the best of ideas because it is primarily programmed by humans which can introduce unintended bias from the very outset. This is because AI systems are heavily dependent on the type and quality of data supplied, and if such data is itself inherently biased it can lead to worse consequences. For instance, in the United States, populations that have historically been disproportionately targeted by law enforcement, particularly the low-income and minority communities if faced with such predetermined law, it will further lead to their discrimination instead of bringing justice. [5]
Another issue that may arise is the identification of correlations. AI is largely based on identifying correlations within data sets. One famous example could be the correlation between low income and a person’s propensity towards crime. There may not be a direct correlation between the two, however, at times poverty may create conditions that make the commission of crimes more likely. Such correlations that may exist cannot be said to be true for every case, and therefore could lead to false positives in crime data, resulting in penalties that may be too severe or too lenient.
Judges are free to accept or ignore the recommendations from the AI. But this might not be possible if humans were completely removed from the process. While AI can be conveniently used to assist in doing jobs like gathering evidence or estimating the likelihood of recidivism, but it should not be preferred while in making final judgments and sentencing decisions.
Technical Challenges
As discussed above, machine learning is based on the predictive algorithms and the machine acts on the basis of available data. One of the major hurdles that India would face in using AI is that the technology and machines currently being used are outdated, the data is often not complete, and unless a huge chunk of reliable data is provided to the machine it will not be able to perform effectively.
There is currently a scarcity of open access to judicial information and datasets, a fact that has been thoroughly discussed and critiqued by Vidhi’s JALDI mission. It becomes crucial that all existing information, and information which will be collected in the future, be archived into readily available datasets, in compliance with recognized principles of open access to data. The computerization of Indian judiciary under the e-courts mission mode project has led to a robust framework to facilitate open access to information captured by Indian courts. However, in the absence of overarching open data policy, the information collected remains scattered and haphazard.
The transformation of judiciary via emerging technologies will bring new challenges. The usual debate of transparency and accountability will then be coupled with the want of future-proofed open data policy. Moreover, it is critical to acknowledge the need for strong data privacy laws that can provide safeguard against increasing cyber threats in the digital age.
Conclusion
The present crisis has triggered the shift from a traditional work environment to a digital one. Although the technology isn’t advanced enough to fully understand the human nuances yet, it cannot be denied that technology and machine learning is the way forward. It’s time that issues of inherent bias or privacy be addressed before resorting to the technology.
Endnotes
[1] Over 3.5 Crore Cases Pending Across Courts in India, Little Change in Numbers Since 2014, The Wire, (Nov. 27, 2019), https://thewire.in/law/pending-court-cases.
[2] Vivek Kumar, Union Budget 2019: What Is It Means for AI and Big Data Industry in India, Analytics Insight, (July 08, 2019), https://www.analyticsinsight.net/union-budget-2019-means-ai-big-data-industry-india/.
[3] Surabhi Agarwal, Technology industry cheers Budget’s focus on AI, ML; asks for revival of SEZ policy, The Economic Times, (Feb 01, 2020), https://economictimes.indiatimes.com/tech/ites/technology-industry-cheers-budgets-focus-on-ai-ml-asks-for-revival-of-sez-policy/articleshow/73845732.cms?from=mdr.
[4] Report of committee – A on platforms and data on Artificial Intelligence, MeitY, July 2019, https://www.meity.gov.in/writereaddata/files/Committes_A-Report_on_Platforms.pdf.; Report of committee – B on leveraging Artificial Intelligence for identifying national missions in key sectors, MeitY, July 2019, https://www.meity.gov.in/writereaddata/files/Committes_B-Report-on-Key-Sector.pdf; Report of committee – C on mapping technological capabilities, key policy enablers required across sectors, skilling, reskill, MeitY, July 2019, https://www.meity.gov.in/writereaddata/files/Committes_C-Report-on_RnD.pdf.; Report of committee – D on cyber security, safety, legal and ethical issues, MeitY, July 2019, https://www.meity.gov.in/writereaddata/files/Committes_D-Cyber-n-Legal-and-Ethical.pdf..
[5] Karen Hao, AI is sending people to jail and getting it wrong, MIT Technology Review, (Jan. 21, 2019), https://www.technologyreview.com/2019/01/21/137783/algorithms-criminal-justice-ai/.