Open Access Research Article
FROM DATA TO VERDICT NAVIGATING AI.S GROWTH & BLEMISH IN THE LEGAL SYSTEM
Published Paper
Article Details
FROM DATA
TO VERDICT: NAVIGATING AI'S GROWTH & BLEMISH IN THE LEGAL SYSTEM
AUTHORED BY -
FAHEEM KHAN
Abstract
The legal industry is
experiencing a surge in the use of artificial intelligence. AI has the
potential to automate and improve efficiency through automation, but the rapid
growth of AI is raising important ethical questions, such as those of job
placement or the disruption of academic experience. The legal industry is
experiencing a surge in the use of AI, with large language models such as
ChatGPT making a huge impact. These models are able to generate text that
mimics human writing, allowing them to be used for summarizing legal docs and
predicting case outcomes. AI is being employed for contract composition,
predicting legal outcomes and even providing judges with advice regarding
sentencing and bail decisions. This article focuses on the unprecedented growth
of AI in legal sector and how AI systems are used when dealing with policy
questions, ethical dilemmas, or legal requirements.
Keywords: legal industry,
artificial intelligence, automation, experience, language model, ChatGPT,
ethical dilemmas
Introduction
Advancement in artificial
intelligence has been rapidly picking up steam lately. Just over the past few
months alone, the release of chatbots seems to have shaken up the world on many
levels, offering exciting new opportunities while also raising some potential
red flags. So, while the potential to automate and improve efficiency through
AI is undeniable, the rapid growth of artificial intelligence is raising
important ethical questions, such as those of job placement or the disruption
of academic experience.
The legal industry is
experiencing a surge in the use of AI, with large language models such as GPT-3
making a huge impact. These models are able to generate text that mimics human
writing, allowing them to be used for summarizing legal documents and predicting
case outcomes. This trend is expected to continue in the future.
Several factors are
contributing to this increase in the utilization of big language models in the
legal sector. These include regular advancements and enhancements of these
models, making them more accurate and complex over time. Furthermore, there's a
growing industrial demand for automation and efficiency which has made these
language models even more beneficial for legal professionals. Additionally, the
need for expedited service is increasing with legal practitioners looking for
ways to speed up their workflows. The use of large language models can help by
automating certain processes while also improving research and analysis.
Machine
Learning Progress & Chat GPT
Machine learning is the
ability of a computer to adjust its operations when new data is fed in. In
order to anticipate future behaviour, predictive models are run on computers,
which use historical data. Therefore, data is essential for machine learning;
the more data available to it, the better it functions. However, the accuracy
and usefulness of the outcomes can be affected by how high quality the data is,
how it is inputted into the system, and how it has been trained to analyze that
data. Additionally, even if the data was correct to begin with, any biases
introduced during its "education" by humans might not have been
intentional.
Consequently, an
algorithm's accuracy relies on both its programming and its data inputs. This
leads us to a legally complex conclusion: the programming behind an algorithm
may be examined and looked into if questions arise about its outcomes.
Unfortunately, it might not be possible to completely identify and analyze all
the data used by the algorithm due to the vast amount of information available
online.
ChatGPT is a computer
program from OpenAI which launched in November 2022. You can think of it as
having a conversation with a robot that knows all information prior to 2021.
GPT stands for Generative Pre-trained Transformer, meaning it was trained on a
massive text dataset and can generate new responses based on what it has
learned. One great thing about GPT is that it can summarize and reinterpret
text so that even a 10-year-old could understand.
Unprecedented
Growth
The base of technological
improvement is notoriously unprecedented for years. Economists said industrial workers could be
reduced to a set of rules that could be computerized. They assumed that
professionals like lawyers are safe because their work was wrapped in language.
But advancement in AI has proven that wrong. Technology has unlocked the
routine task of shifting through documents & looking for relevant pages.
Now many law firms have taken initiatives to understand and use this emerging
technology.
India's legal tech
start-ups have raised $57 million in the last decade with more than 650 start-ups
and the sector worth $1.3 billion, less than 1% of the US market. It's
considered the second biggest in the world and is on the rise due to digital
code, proceedings and legal costs that take up about 0.48% of GDP. Funding for
this sector began in the year 2013 with 32 start-ups receiving 70 rounds of
investment until June 2022. The year 2021 saw the most investment yet with
nearly $20 million allocated to this sector. Although at least stage funding
dominated these operations until recently, growth stage investments have been
increasing as the market becomes more profitable.[1]
These start-ups focus on
specific areas of law and their software is trained by human experts in order
for it to learn and advance. When Alexander Hudek, a computer scientist and
chief technological officer of Kira Systems, started his project of automating
contract reviews back in 2011, he expected it would take him 4 months to tweak
standard algorithms. But to his surprise it took him two and a half years
instead to refine the software so that it could readily identify legal
concepts.[2] Kira Systems is a Canadian machine learning
program which has been reported by its clients about the reduction in lawyer
time required for contract review between 20 to 60% due to the efficiency of
the program. Despite this, however, human scrutiny is still required to finesse
the job said Noah Waisberg, Chief executive of Kira.[3]
Courtroom
Influence
Blue-collar occupations
are increasingly being optimized by AI, but it is now also transforming
white-collar vocations that were once thought to be safe from automation. These
superhuman capabilities are leading to complete workplace overhauls, with human
co-workers often being supplemented or even replaced by AI.
Even in more established
justice systems, access to legal assistance remains a significant challenge,
with only 46% of the population having access to the court system and backlogs
can be overwhelming. To resolve this issue, technology is being used to move
judicial proceedings online and create an accessible service rather than a
physical destination.[4]
AI can currently be
utilized for contract review, document discovery in litigation cases, and legal
research; however, it is more likely to support lawyers instead of replacing
them in the short run. Furthermore, AI is being employed for contract
composition, predicting legal outcomes and even providing judges with advice
regarding sentencing and bail decisions.
The Covid-19 pandemic
shed light on the necessity for the digitalization of India's legal system and
greatly accelerated the process. This began with the initiation of the 'eCourt
Project', a project under the National e-Governance Plan that was announced in
2007. Artificial Intelligence (AI) is currently being utilized by India's
Supreme Court to transcribe its proceedings in real time, using a natural
language processing technology known as 'Teres' developed by Bangalore-based
Nomology Technologies Private Limited.[5] The transcribed text will be published every
evening on their website, and this is a big step towards making judicial
proceedings more transparent. Moreover, last year saw the introduction of live
streaming for proceedings held before a Constitution bench.
The Indian judiciary has
additionally developed an AI-driven portal called ‘SUPACE’, Supreme Court
Portal for Assistance in Courts Efficiency, as well as launched an AI
translation tool called ‘SUVAS’, Supreme Court Vidhik Anuvaad Software, to
facilitate translations from English to Marathi, Hindi, Bengali, Punjabi,
Kannada, Tamil, Telugu, Gujarati and Malayalam.[6]
Recently, the Punjab
& Haryana High Court became the first court to use Chat GPT technology to
decide a bail plea of an accused and eventually rejected the petition. Justice
Anoop Chitkara requested a response from Chat GPT while hearing the bail
application of an accused arrested in June 2020 for alleged rioting, criminal
intimidation, murder and criminal conspiracy. The judge assessed the reply from
Chat GPT and denied the bail plea using his previous experiences and decisions
as grounds. He said that “To inflict death is cruel in itself, but if cruelty
leads to death, then the situation changes. When a physical assault is
committed in a brutal manner, the parameters of bail also shift”, stressing
that any reference or observation to Chat GPT was only intended to provide
further insight into bail jurisprudence related to cruelty.[7]
Commenting on this
judgment's use of chat GPT, many high court lawyers were of the opinion that it
can have both positive and negative effects, which will be determined in time.
Some regarded the use of AI as a tool rather than a threat, stating that it is
too early to declare that AI could threaten lawyers’ jobs due to its initial
stage. For example, when computers arrived, people feared they would lose their
jobs, yet now these machines are part of IT, creating more employment
opportunities than before.
AI &
Legal Professionals: Collaboration or Competition
Artificial Intelligence
has disrupted a variety of industries, including the legal field. While some in
the legal profession view AI with concern for their job security, others view
it as a way to improve their work. In practice, partnering with AI tools and
lawyers can lead to better client results, but this requires learning to use
new technologies.
Advocates of AI point out
it can take on large and complex data sets more quickly than humans, which help
attorneys make accurate judgements. Machines are now capable of completing
tasks previously thought of as the job of a lawyer analyzing large amounts of
documents related to dispute resolution and even drafting documents. We're even
beginning to see machines predict the outcomes of disputes!
The lawyers of tomorrow
will be those who build systems able to solve clients' problems. They'll
include experts in legal knowledge engineering, legal risk management, system
development and design thinking. The legal sector is going through a transition
similar to what other industries have experienced which makes sense because it
is highly reliant on documentation, thus making it very well-suited for
technology's benefits.
An example of this is
India's Cyril Amarchand Mangaldas, which licensed Kira to carry out multiple
tasks simultaneously and also to analyze legal papers, pin-point important
points and extract regulations from various documents, ultimately saving
considerable time and resources.[8]
On the flip side, some
legal professionals are concerned that AI could eventually take over the
profession. There is fear that as AI technology progresses, it will be able to
manage more intricate legal duties, such as creating legal documents and
arguing in court. If this happens, lawyers and paralegals could find themselves
without jobs.
Nevertheless, AI cannot
substitute for the human opinion, resourcefulness, and empathy that are key
parts of the legal profession. Lawyers must interpret the law and apply it to
each particular case with an understanding of the law and the capacity to
examine complex conditions. AI can help attorneys carry out these tasks, but
not replace them.[9] Plus,
the legal industry is heavily regulated, with many legal tasks requiring a
license to practice law; this something AI cannot acquire.
But eventually, how AI
and legal professionals collaborate will depend on how technology is developed
and employed. If AI is employed as an instrument to enhance human skillset and
improve quality of legal services then it might be beneficial for the
profession. But if AI replaces human professionals entirely then it could
result in substantial job losses and a decrease in quality of legal services.
Potential
Pitfalls
AI could potentially
streamline the legal process, including research and decision-making. However,
this technology is not capable of replacing human judgment entirely within the
legal field. AI has the potential to contain embedded bias within its data and
presents difficulties in terms of comprehending its decisions in a format understandable
to humans. Therefore, these issues must be addressed before AI can be reliably
used in a legal context.
AI has been called a
potential revolution across all human affairs, with some believing it could be
the most influential invention in history. It is essential to consider the
implications of how AI systems are used when dealing with policy questions,
ethical dilemmas, or legal requirements. Many experts in this area have spoken
out against using AI in court proceedings as they do not possess the intellect
needed to solve complex legal issues, let alone any moral or ethical
considerations.
For instance, using AI to
create legal contracts requires a developer to collect labelled data on
different versions of contract texts, enabling the AI to learn how to construct
an acceptable contract. However, the implementation of a contract is usually
specific to its context, varies by jurisdiction, and is subject to an
ever-changing body of law. Furthermore, because most contracts are never
litigated or enforced in court, their specific terms remain confidential
between the parties. This raises the concern that AI-generated contracts could
amplify both good and bad legal work.[10]
Therefore, it's uncertain how much better AI contract writers can become in the
near future since they lack the domain experience and linguistic accuracy for
autonomous operation. Despite this, these tools can still be useful when
drafting language; however, it is necessary for a human professional to review
the output before it is employed.
Similar to any new
technology, there are regulatory issues too that need to be addressed in order
for it to be safely and ethically used. Of the many challenges presented by AI
in this industry, transparency and explainability are two of the most
important. Many AI systems don't make it clear how they arrive at their
decisions, which doesn't work well for the legal profession where judgements
need to be supported and explained. Without this clarity, there is a risk of
making decisions which can appear biased or hard to understand, raising
questions about justice and due process.[11]
To avoid this issue, legal practitioners must ensure that their AI systems are
transparent and able to detail the system's function, its data sources, as well
as the reasoning behind its decisions.
Vidhi Center for Legal
Policy, an independent think-tank, has recently highlighted the need for a
comprehensive legal, regulatory and ethical framework to be implemented before
AI can be integrated into the justice system.[12]
This is due to concerning bias found in AI technologies over the past few
years. Moreover, data security and privacy must also be taken into
consideration. Legal professionals must make sure that AI systems meet
regulations like the GDPR in Europe and that only authorized personnel can
access the data. Under GDPR, people have the right to know what kind of
personal data is being used, the right to get their data erased, and the right
to object to automated decision-making which might be utilized in AI systems.[13]
So that means AI systems need to be transparent and easily explained.
Additionally, existing anti-discrimination laws should be transferred onto
digital platforms so as to protect consumers from any potential unfairness and
to ensure trust in these systems.
In case of Chat GPT,
while it may accomplish a lot, it still has a few of restrictions. You can
request it to generate any kind of content, but it's your responsibility to
take the output and rephrase, correct, and arrange it. This chatbot is like an
attentive personal assistant who will do as you ask, but you must know how to
phrase the request and make use of the output. Experts seem to agree that AI
will not be able to generate entirely original ideas on its own anytime soon.
The best way to stay competitive is to nurture creativity; keeping a creative
mindset while using AI technology to increase productivity will make you
invaluable.
James Yoon, a seasoned
attorney from Palo Alto, California, reflects on the time in 1999 when a
significant patent matter required three partners, five associates, and four
paralegals. Due to limited legal resources and technology that performs
essential work, a similar case today only requires one partner, two associates,
and one paralegal. To plan for patent litigation, Mr. Yoon has been using
software programmes like Lex Machina and Ravel Law. To create profiles of
attorneys and forecast results, these programmes comb through court judgements
and filing data. Similar to how analysts of baseball and football examine the
tendencies of players on various teams, similar analysis is done for the legal
profession.[14]
However, modern
technology still only complements human labour; it has not yet fully supplanted
it. Mr. Yoon spends the majority of his time using his judgement, creativity,
and strategy, which are tasks that machines cannot currently perform. His
billing rate, which was $400 per hour in 1999 when he was 49 years old, has
significantly climbed to $1,100 per hour now, demonstrating the value of
expertise. He thinks that technology is turning more and more labour into
routine chores, but for the time being, clients are still willing to pay more
for human experience than for routine jobs.
Conclusion
AI is a reflection of us
and can show us the flaws in our own human reasoning. It's possible that when
algorithms make predictions, it could be because of our own bias, as the codes
and algorithms that power AI may be built with certain values in mind. That's
why it might be morally wrong to let machines decide on matters of personal
liberty.
However, AI has already
been used in due diligence and case management, although they are still far
from reliable - it could be dangerous to put too much trust in machines due to
something called automation bias - the idea that we might overestimate the
capabilities of AI simply because it is automated. So even though AI can be
powerful, it would be wise to remember that it is still not as advanced as we
may think.
[1] Aryaman Gupta, Legal tech start-ups have raised $57
mn in funding since 2013, Business Standard (Dec. 22, 2022), https://www.business-standard.com/article/companies/legal-tech-start-ups-raised-57-million-in-funding-since-2013-says-report-122122100918_1.html.
[2] Steve Lohr, A.I. Is Doing Legal Work. But It Won’t
Replace Lawyers, Yet., The New York Times (Mar. 19, 2017), https://www.nytimes.com/2017/03/19/technology/lawyers-artificial-intelligence.html.
[3] Kathryn D. Betts And Kyle R Jaep, The Dawn of Fully
Automated Contract Drafting: Machine Learning Breathes New Life Into a
Decades-Old Promise, 15 DUKE LAW & TECHNOLOGY REVIEW 9-12 (2017), https://scholarship.law.duke.edu/cgi/viewcontent.cgi?article=1306&context=dltr
[4] Bernard Marr, The Future of Lawyers: Legal Tech, AI,
Big Data And Online Courts, Forbes (Jan. 17, 2020), https://www.forbes.com/sites/bernardmarr/2020/01/17/the-future-of-lawyers-legal-tech-ai-big-data-and-online-courts/?sh=69484622f8c4.
[7] Jagpreet Singh Sandhu, In a first, Punjab and Haryana
HC turns to ChatGPT for view on bail in murder case, The Indian Express (Mar.
28, 2023), https://indianexpress.com/article/cities/chandigarh/in-a-first-hc-turns-to-chatgpt-for-view-on-bail-in-murder-case-8522544/oceedings-564632.
[8] Shreeja Sen, Cyril Amarchand Mangaldas in pact with
software solutions firm Kira Systems, Live Mint (Jan. 31, 2017), https://www.livemint.com/Companies/nJErdQk632tBn4MBENb9aM/Cyril-Amarchand-Mangaldas-in-pact-with-software-solutions-fi.html.
[9]
Rohit Narain, AI for Law: Can AI Replace Lawyers?, DataToBiz (Oct. 29, 2021), https://www.datatobiz.com/blog/ai-for-law/.
[10]
Heiko Hotz, How to Set up a Machine Learning Model for Legal Contract Review,
Towards Data Science (June 4, 2021), https://towardsdatascience.com/how-to-set-up-a-machine-learning-model-for-legal-contract-review-fe3b48b05a0e.
[11]
Nagadivya Balasubramaniam, Transparency and Explainability of AI Systems: From
Ethical Guidelines to Requirements, 159 Information and Software Technology 1-3
(2023).
[12] Pritam Bordoloi, The Power & Pitfalls of AI in
Indian Justice system, Analytics India Magazine (July 25, 2022), https://analyticsindiamag.com/the-power-pitfalls-of-ai-in-indian-justice-system/.
[13] Avantika Bhandari, The Impact of the GDPR on
Artificial Intelligence, Montreal Ethics (Feb. 20, 2022), https://montrealethics.ai/the-impact-of-the-gdpr-on-artificial-intelligence/.
[14] Steve Lohr, A.I. Is Doing Legal Work. But It Won’t
Replace Lawyers, Yet., The New York Times (Mar. 19, 2017), https://www.nytimes.com/2017/03/19/technology/lawyers-artificial-intelligence.html.
Article Information
FROM DATA TO VERDICT NAVIGATING AI.S GROWTH & BLEMISH IN THE LEGAL SYSTEM
Authors: FAHEEM KHAN
- Journal IJLRA
- ISSN 2582-6433
- Published 2023/10/31
- Issue 7
About Journal
International Journal for Legal Research and Analysis
- Abbreviation IJLRA
- ISSN 2582-6433
- Access Open Access
- License CC 4.0
All research articles published in International Journal for Legal Research and Analysis are open access and available to read, download and share, subject to proper citation of the original work.
Disclaimer: The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of International Journal for Legal Research and Analysis.