From creating dazzling digital art to diagnosing diseases with unparalleled accuracy, artificial intelligence ("AI") is pushing the boundaries of what is considered possible across numerous industries. Now, AI stands on the brink of revolutionizing dispute resolution, promising faster and more efficient processes that could reshape how we resolve conflicts. However, this potential brings significant practical, philosophical, and ethical challenges. The Chinese Room argument encapsulates a key concern: just as someone following English instructions to produce Chinese characters without understanding the language, AI may only simulate intelligence, adhering to pre-set rules without true comprehension. This limitation is crucial in dispute resolution, where human judgment, empathy, and ethical reasoning are indispensable. As we move towards an AI-driven era in the legal field, it is important to acknowledge that while AI can enhance efficiency, it cannot replace the uniquely human elements of justice. AI may be a powerful assistive tool in dispute resolution, transforming tasks like document production, legal research, and drafting. But as we harness AI's capabilities, we must ask: can we really trust machines to comprehend and deliver justice in its truest form?
I. Introduction
In a world where justice is often slow and expensive, the promise of artificial intelligence to revolutionize dispute resolution is both tantalizing and terrifying. Over the last decade, the world has seen a paradigm shift in the field of technology and machine learning. The arrival of Artificial Intelligence ("AI") applications such as Chat-GPT, has drastically changed the market of machine learning and AI. Today, AI has been integrated in almost every industry and is being widely used for various operations. Dispute resolution is one area in which the use of AI can potentially lead to a total disruption. There are a myriad of ways in which AI can prove its use case for dispute resolution. However, as AI continues to advance and permeate the legal field, it is critical to examine both the potential benefits and the risks it poses for dispute resolution.
Chinese Room Problem
At the outset, it is worth exploring some of the deeper philosophical questions that arise when we examine the nature of artificial intelligence. One of the most famous thought experiments in the philosophy of AI is the Chinese Room argument. Imagine a person who does not understand Chinese sitting alone in a room. They are given a set of instructions in English for manipulating Chinese characters, and then provided with Chinese characters as input. By following the instructions, they are able to produce Chinese characters as output that are indistinguishable from those a native Chinese speaker would give. To an outside observer, it appears that the person in the room understands Chinese. But in reality, they are simply following a set of rules without any true comprehension. It is argued that this thought experiment demonstrates that even if a computer program can simulate intelligent behavior, it does not necessarily possess true understanding or consciousness. In other words, the appearance of intelligence does not equal genuine intelligence.
This above philosophical conundrum has significant implications for how we think about AI and its potential applications in dispute resolution. While AI systems may be able to analyze vast amounts of data, identify patterns, and generate solutions that mimic those a human expert would provide, can we say that they truly understand the nuances and complexities of legal disputes? Or are they, like the person in the Chinese Room, simply following a set of pre-programmed instructions without genuine comprehension?
Currently, it appears that AI may have the potential to make dispute resolution faster, cheaper, and more efficient. However, it may come with its own set of limitations particularly where human judgment and empathy is needed. Accordingly, as AI continues to advance and demonstrate its value in the dispute resolution process, it is important to recognize that the technology is likely to serve as a powerful tool rather than a complete replacement for human mediators and arbitrators in the coming years.
II .Exploring AI's Role in Dispute Resolution
The pandemic-era already taught us that we could take the dispute resolution process entirely online with the help of Online Dispute Resolution ("ODR") platforms. Now integrating ODR with AI has become much easier, thereby further enhancing efficiency and costs effectiveness. AI is showing signs of huge potential in areas like:
While there is general consensus on the benefits of AI as an assistive tool, its potential to replace human judges and arbitrators remains controversial. A survey by Queen Mary University found that 78% of survey respondents favored the integration of AI into dispute resolution processes. However, legal professionals across jurisdictions are sceptical about the idea of replacing judges and arbitrators with AI.
Further, in the absence of specific regulations, ethical concerns surrounding confidentiality and attorney-client privilege may arise even while using AI as an assistive tool. The efficacy of AI in multicultural contexts can be constrained, as it may struggle to fully interpret and capture cultural subtleties.
Thus, the use of AI in the field of dispute resolution could be categorized broadly into two categories:
In this article, an attempt has been made to examine the pros and cons of both aspects, i.e. AI as an assistive tool and AI as a decision maker.
(A) AI as an assistive tool
A well-structured and trained AI can wear multiple hats. It can be used as a virtual assistant for the entire proceedings helping the adjudicators and counsel in locating relevant documents and pleadings. It is also helping counsels, attorneys, arbitrators, and other adjudicators in transcription services and provides various tools for foreign language interpretation which ensures more clarity and accessibility in the proceedings.
AI-powered ODR platforms, such as the American Arbitration Association's Modria Resolution Centre, offer structured environments where disputes can be resolved remotely. Additionally, in the present era, AI is significantly helping the advocates/counsel in the document management process. Even various experts have predicated AI as an efficient document manager wherein AI can substitute a physical assistant like junior advocates, interns, paralegals, or for that matter tribunal secretaries. AI also aids in case prediction and analytics, providing valuable insights by analyzing past decisions to predict outcomes and inform legal strategies. For instance, there are various computer programs which are presently available to help attorneys in analysing the other side's written submissions and to provide relevant case law that was omitted therein or has been rendered since. Although the technology's potential in complex decision-making is still developing, AI's current use in case management like automating scheduling, notifications, and tracking, demonstrates its capability to minimize human error and administrative burdens.
Another form of an efficient use of AI is Natural Language Processing ("NLP"), having the capability of reading, interpreting, and communicating in human natural language. The NLP-enabled AI can be used to do tasks including the review of documents and proof reading of the documents presented by the client in
minutes. Further, a more desirable situation would be, to modify NLP in such a manner that the lawyers could simply feed in all the client-related data and pose questions to the AI that are strategic to the case. If NLP can respond accurately by going through the documents and without hallucinations, complexity in e-discovery will surely become a thing of the past.
Key Challenges for using AI as an assistive tool
However, while AI offers substantial benefits, it also poses challenges such as ensuring data privacy, preventing algorithmic bias, and maintaining essential human oversight in legal decisions. The risks associated with the usage of AI in dispute resolution proceedings can be broadly categorized as -
iii. Cultural bias - in international cross-border dispute resolution, which involves multiple legal cultural frameworks, AI algorithms face unique challenges, particularly concerning cultural biases. An AI tool trained primarily on US laws and legal precedents might inadequately apply the same framework when interpreting UK laws, despite fundamental differences between the two countries' legal systems. This scenario was evidenced when an AI Assistant, initially trained on US court decisions, failed to accurately interpret documents governed by UK laws during a dispute resolution process. Such instances highlight the AI tool's susceptibility to the "garbage in, garbage out" phenomenon, where the quality of AI outputs directly correlates with the quality of its training data.
(B) AI as a decision maker
One of the most burning questions is whether AI can function as an active decision-maker in dispute resolution. Various studies have demonstrated the accuracy of AI in predicting court decisions and resolving disputes. A 2017 study showed that a machine learning model predicted US Supreme Court decisions with 70.2% accuracy.The Guangzhou Arbitration Commission in China developed an AI arbitral assistant that successfully resolved a dispute, claiming a fourfold efficiency improvement. After hearing the dispute, the assistant mentioned "Today's hearing has come to an end. I am currently analyzing the trial data and the ruling opinion will be sent to the arbitration tribunal via email in 5 minutes". In Canada, a robot mediator settled a three-month-long dispute in just one hour.
However, these case studies were based on set patterns of data. It remains to be seen whether AI can handle complex moral and ethical dilemmas.
An interesting question that can arise would be in relation to the validity of the expert report generated by AI in an arbitration proceeding. Unlike a purely assistive role, where AI supports the expert, this scenario involves the AI itself determining and opining on the correct approach. With the advent of reinforcement learning which trains software to make decisions to achieve the most optimal results, AI can be made competent to provide expert reports on complex technical subject matters including technology and construction disputes. However, the validity of such a report would face a lot of scrutiny as experts are still categorized as persons in almost every jurisdiction. Further, there would be a great setback with respect to procedural equality as the opposite party would not be able to cross-examine the AI expert. However, in the absence of any specific legal framework, using AI by an expert for better expert reports looks like a much safer option.
The decision-making power and winning probability assessment power of an AI are most useful in the context of third-party funding arrangements. Third-party funders majorly fund the claim of parties in lieu of some part of the award amount in case of a favorable outcome. With respect to the assessment of the claims of the parties, third-party funders majorly take the assistance of lawyers to determine the winning probability,
however, such opinion is not always accurate as there could be bias in the probability provided by the lawyers. AI in these situations can provide an unbiased winning probability to the third-party funders on the basis of precedents, which can effectively help in better decision-making by the funders.
Key challenges to AI as a decision maker
Using AI as a decision-maker in cross-border dispute resolution mechanisms such as arbitration introduces a large number of issues including the following -
III.Navigating the Future: Trends and Challenges Ahead
The dispute resolution industry has been slow to adopt new technologies largely attributable to the conservative inclinations of lawyers and adjudicators. However, the COVID-19 pandemic has accelerated the integration of technology into the dispute resolution process - from online filings to virtual hearings - thereby facilitating increased utilisation of AI technology. Arbitrators are increasingly substituting human tribunal secretaries with AI assistants to improve efficiency and cost-effectiveness, while lawyers are progressively leveraging AI to optimise their practice. However, the legal community often views AI as an inexperienced intern whose work cannot be trusted, due to obstacles such as hallucinations and confidentiality issues. As the industry develops, these problems are expected to be addressed by more advanced AI programs.
The biggest drawback of AI as a decision-maker or assistant is the lack of trust among the legal community and parties involved. To generate trust, there is an urgent need for regulating AI in the legal space. Institutions should test the accuracy of AI technology and provide accreditation to AI algorithms. Regulations can also address concerns such as hallucinations, confidentiality, and bias. Before integrating AI technology with clients' confidential data, it should be mandatory to seek consent of such clients and adhere to their preferences for employing specifically accredited AI algorithms to handle their disputes.
There is also uncertainty regarding how enforcement courts would react to AI-generated decisions and awards. However, AI can help arbitrators or judges reduce human bias by comparing their judgments with AI-generated outcomes. With AI assistance, there may be less need for three-arbitrator panels, as AI can help neutralise individual arbitrator bias. In the present scenario, AI poses a benefit to the legal fraternity by making the dispute resolution process more cost-effective and time-efficient when used as an assistive tool. However, completely replacing judges and arbitrators with robots without an established legal framework or regulation would be premature.