Authors: Kunal Verma, Waheb Hussaini and Aniruddha Das
Introduction:
Artificial Intelligence or AI is no longer an emerging innovation – it is now a transformative tool which has a potential of being considered the next industrial revolution. AI in legal space is not far behind! For legal practitioners AI is a productivity enhancing tool which can be used to analyze case-papers, articulate strategies, creating template documents, improve research, structure precise arguments, draft complex agreements and much more. In no time, it will be the team member with no threat of attrition. In Alternate Dispute Resolution (“ADR”), legal practitioners, mediators and arbitrators are using AI to enhance efficiency and effectiveness. AI has shown immense promise in the area of precisely analyzing data in matter of minutes, which not only reduces the burden of spending several billable hours of the resources but also provides the team more space for strategic thinking. While Artificial Intelligence has numerous benefits and appears to be the cutting edge technology, however we will be doing injustice to our human intellect to blindly implement AI in our systems without recognizing and understanding the limitations and risks that it comes with.
Risks:
Various studies and user data have exposed some known risks and shortcoming which legal practitioners must address before integrating AI in their systems, as these pitfalls have professional and ethical consequences.
i. Data security and confidentiality
Confidentiality is the foundation of trust shared between an attorney and client. Imagine, if highly sensitive documents relating to mergers and acquisitions or a hotly contested divorce are leaked. The consequences would be devastating and far reaching to say the least – not just for the law firm but also for the client. Using AI makes this risk real. Uploading documents containing highly sensitive and confidential information on AI platforms has the potential of exposing this data to unauthorized access, data breaches and misuse. In fact, the concern is not just restricted to confidentiality of the content, but also with regard to further use of such data for the training and upgrading the AI system. Water tight security for storage and access of data, therefore, is a crucial concern while integrating AI into work culture.
ii. Limitation of Technology
- Hallucination
In context of AI, Hallucination means creation of seemingly credible but factually false and misleading information, which may include generation of fake case-laws or statutes. AI relies on large language model (LLM) which, with the help of prompt
engineering, can be finetuned to perform specific tasks including doing targeted research. However, such models come with inaccuracies present in the data they are trained in. This would mean that a law practitioner, basing their research only on AI, may become a victim of hallucinated data and may run a risk of misrepresenting facts and information and even law in some cases. We all are aware of the unfortunate incident that happened before US District Court where a law firm faced sanctions for filing a motion referring to hallucinated information. The motion relied upon 8 AI- generated cases which did not exist. - Systemic bias present in the dataset
Another significant risk of AI legal chat bots is that of ‘systemic bias’ or ‘algorithmic discrimination’. The dataset relied upon by AI may have inherent complexities like
societal inequalities and other biases, which may potentially lead to unfair outcomes in dispute resolution contexts and may affect application of legal principles to relevant facts. If this data reflects historical prejudices or is unrepresentative of the broader population, the AI’s decisions may perpetuate or amplify existing biases leading to
unfair treatment of an individual or group. Hence use of AI technology with unreliable data, may have serious ethical concerns involving the principles of neutrality and impartiality in ADR process. - AI cannot replicate ‘Human Factor’
Even though Artificial Intelligence has become a central part of our work culture, human judgment, intuition, empathy and even discretion is still indispensable. Human emotion and human intelligence are critical components in adjudicating variety of matters. In ADR, fostering trust and confidence in the process is a key factor. In mediation, especially in emotionally charged matters, the resolution may go beyond purely legal considerations and in such situations use of Artificial Intelligence will be limited. Similarly, in arbitrations, cross-examination of a witness involves evaluation not only of the answers to the questions, but also on the conduct, context, tone and tenor of the witness. Delegating these aspects to AI may lead to flawed interpretations and undesirable results. - Transparency and accountability
Use of information in a fair, equitable, and transparent manner is
foundation for ethical deployment of AI system. The opacity of AI, use of
complex algorithms, lack of transparency in the decision-making process
can undermine the foundational principles of accountability and fairness in
the legal system. Clarity on data policy, algorithm transparency and open
disclosure on limitations employed by the AI system, will generate
awareness and ensure the attorneys adopt necessary precaution in usage of
Artificial Intelligence.
Mitigation of Risk:
To mitigate the risks involved in integrating AI systems, a right balance needs to be maintained between usage of AI to drive efficiency and productive on one hand and the necessity to ensuring compliance with regulatory regime and protecting clients confidentiality on the other hand. Some of the simple best practices are:
i. Complete disclosure and consent
Ethical use and integration of AI is a need of hour to enforce trust and foster confidence in the dispute resolution process. In order to being compliant with the fiduciary duty every lawyer should adopt the practice of complete disclosure to their clients regarding the use of AI technologies. In ADR, the parties must be clearly informed about any planned use of AI, the advantages and the limitations involved. A disclosure at the very initial stage must be adopted as an industry practice. This will empower the clients to make an informed decision and give consent for use the AI technology.
ii. Human oversight to maintain control over AI
Although AI has made advancements towards data security and confidentiality, however, it is at a very nascent stage and is mostly unregulated. The most effective and imperative safeguard while integrating AI into our work culture is and will remain, continued presence of active human supervision and control.
In order to supervise effectively, legal practitioners should minutely go through the fine print to understand the manner and extent of use of confidential data keyed in to the AI platform. Further, professionals should conduct periodic evaluations of the security measures adopted by the AI platforms integrated in their systems. Furthermore, to avoid a risk of hallucinated data and potential inaccuracies which may lead to misrepresentation or misinterpreting the law, a user must undertake reasonable inquiry into the output produced by AI and ensure the reliance on research is not without cross verification prior to presenting it before the Court or in any proceeding.
iii. Development and implementation of regulatory framework
With explosive growth of AI technologies, there is a grave urgency for developing and implementing a regulatory framework to govern Artificial Intelligence industry especially in legal practice. As of now, these technologies are evolving and newer practices are being adopted on a daily basis.
Legislation is required for fixing accountability, ensuring compliance with data privacy laws, framework to prohibit AI from creating illegal content and to define parameters concerning security measures. Although some countries and jurisdictions have guidelines and advisories for ethical and responsible usage of AI, still it is largely an unregulated space. Further, considering ADR may involve cross border disputes with overlapping AI technologies, it is advisable for the ADR Institutions to create guidelines to agreeable AI-usage policies by prescribing dos and donts. Internal policies and guidelines will ensure safe, secure, and trustworthy setting for use of AI during the life cycle of the dispute and will ensure a uniform approach in using AI technology.
iv. Creating an AI ready work force – Awareness of Risks and Limitation of AI While it is important to inculcate AI technology, it is equally important to upskill the workforce and make it AI ready. Artificial Intelligence is not meant for replacing humans but to augment and strengthen them. Lawyers and paralegals should not just be aware on how to use the AI technology but should also be made aware as to the extent to which reliance can be based on AI generated results.
Conclusion:
Reliance on AI for any dispute resolution mechanism cannot be unrestricted and unchecked. Work of lawyers, arbitrators and mediators is of great significance as in the process of adjudicating dispute they are also administrating justice which has serious consequences and impact commercial as well as individual rights of parties. Consequently, it is important to implement best practices to use Artificial Intelligence in an ethical manner. The duty falls, not just on the attorneys, but also on the arbitrators and mediators to disclose deployment of Artificial Intelligence during dispute resolution process.
Disclaimer – This paper has been prepared by a human – a real one – without any assistance from or influence by artificial intelligence. Any error or omission is purely human in nature.
Bibliography:
https://link.springer.com/article/10.1007/s44163-024-00121-8
https://www.eve.legal/blogs/disclosing-ai-usage-to-your-clients-best-practices-for-
legal-teams#:~:text=Conclusion,services%20without%20diminishing%20client%20trust.
https://www.icsi.edu/media/webmodules/CSJ/December-2024/17.pdf
This Paper was prepared by the author for panel discussion in the the Alternative Dispute
Resolution Conference 2025, held on 22 & 23 May 2025 which was jointly organised by
LAWASIA and The Law Society of Singapore.
