Author: Rana Sajjad*
This blog post is Part II of a three-part series on “Multi-Intelligence Justice”, a term coined by the author while conceiving the idea of Multi-Intelligence Justice which he has developed into a jurisprudential theory. In part I, the author introduced the idea of Multi-Intelligence Justice and explained the rationale behind it while comparing the strengths of both human and artificial intelligence and discussing how dispensation of justice involved more than logical reasoning and an application of the law to the facts of the case. In this part and the next, the author discusses the broad parameters, key considerations and the practical and normative implications of the Multi-Intelligence Justice framework that he proposes.
| Jurisdiction: |
Topics: |
Part II: The Mechanics of Human-AI Interplay in the Adjudication Process
Multi-Intelligence Justice (MIJ) entails a seamless integration of human intelligence (HI) and artificial intelligence (AI) for dispensation of justice and rendering of arbitral awards in the realm of international arbitration. As proposed in Part I of this blog post series, MIJ’s goal is to have an optimal combination of both HI and AI whereby HI is amplified by the technical capabilities of AI to make dispensation of justice in international arbitration both efficient and efficacious. This human involvement in every aspect of the justice/arbitration process. would, on the one hand, allay concerns of those who believe that AI alone cannot or should not be trusted especially in light of concerns around due process and, in turn, enforcement of arbitral awards. On the other hand, those who believe that humans alone should not be involved in every aspect of the adjudication process would be reassured that the human involvement is not coming at the cost of AI’s obvious benefits including time and cost efficiency. This complementarity would also help us move from the “Human vs AI” narrative to the “Human and AI” one, from competitors to collaborators because ultimately it is the disputing parties’ interests that need to be accorded the highest priority regardless of how dominant the role of HI or AI is. Only then can the justice system, the one that has traditionally existed involving only humans, the hybrid one that is currently being tested with a preeminent human role, and the one MIJ aims to build, can try optimizing timely delivery of justice in its true spirit.
So, how do we ensure that the interplay between HI and AI is optimal? By ensuring that the mechanics of this interplay is uniform, consistent and transparent. To this end, I propose the MIJ framework whose goal would be to provide concrete protocols and guardrails for the interplay between HI and AI, irrespective of the specific role played by HI and AI and whether the roles are performed simultaneously or sequentially, for any given task during the adjudication process. The MIJ framework will, on the one hand, aim to avoid or minimize human errors, shortcomings, biases and, at the same time, enhance human efficiency to dispense justice and render awards with less mental effort and labour. This efficiency and optimal utilization would be accomplished neither at the cost of a sound application of the Arbitrator’s judicial/commercial mind nor at the cost of the innate human traits such as human intuition, compassion, contextual and cultural understanding, and ethical reasoning that are critical for doing justice. By clarifying how and when to use AI, the MIJ framework will help conserve the Arbitrator’s mental, and even emotional, energy and capacity that will, in turn, facilitate an optimal utilization of the human traits and abilities.
Furthermore, unlike a query on a Large Language Model (LLM), this interplay between HI and AI would not take place in a black box. Thus, transparency is another important element of the MIJ framework since it is critical for ensuring MIJ’s procedural integrity and avoid or minimize the possibility of a challenge on due process grounds. An open and clear articulation of the Arbitrator’s role as well as the discretion and the limits within which that discretion is exercised under the MIJ framework would engender a sense of responsibility and cultivate a sense of accountability. If AI’s role is predominant, its basis and scope have to be explained. This would not only enhance transparency of the process but also help ascribe responsibility and liability. Imagine a scenario in which the Arbitrator chooses to disavow responsibility for the legal research or the analysis/reasoning for arriving at a decision which is subsequently challenged by one of the parties to the dispute. In tech speak, if the MIJ framework is to follow the learning loop for iterative growth, we have to know what exactly went wrong and who or rather which intelligence — HI or AI — is responsible for it.
Hence, the MIJ framework has to define the role as well as the limits of the role played by HI and AI, which would, in turn, determine the scope, sequence and manner in which HI and AI are used as well as the mechanics of the interplay between the two. So let us now turn to the mechanics of this integration and the protocols to be prioritized under the MIJ Framework.
The first phase of the adjudication process involves reviewing the case file. In a large number of cases, especially the high-stakes ones in international arbitration, the documents are fairly voluminous. While an Arbitrator may not consider it necessary to read all the documents word for word, there is a possibility that while reviewing the case file, the Arbitrator misses something that is of consequence to the final outcome. To err is human after all (and to hallucinate is artificial, more on that in a bit).
One way of rectifying these human errors is to assign AI the task of reviewing the case file, as a first step. The Arbitrator, while instructing/prompting AI, will give it the background of the case and the context for reviewing the case file. AI’s unmatched computing power will then help separate the proverbial wheat from the chaff in quick time. There is a caveat here though. We have all heard of and may even have experienced how AI hallucinates, that is, provides false or even fabricated information as part of its results. As a safeguard against these hallucinations, the MIJ framework would require that the Arbitrator cross check, at a minimum, the most critical documents, identified by the Arbitrator as well as the specific conclusions drawn from them by AI. While this second review of the case file may raise the question of duplication of efforts, of a potentially time-consuming second review conducted by the Arbitrator, that may undermine the very efficiency for which AI was assigned this task in the first place. But, as part of the iterative growth of this framework, we have to make strategic trade-offs while we learn as we go along. With AI, we run the risk of hallucinations but we trade that for the lightning speed at which it reviews the documents. By the same token, the additional time taken by the Arbitrator for a second review is traded for all the invaluable human traits critical to a fair decision. From the standpoint of responsibility and accountability also, it would be prudent to entrust the Arbitrator the duty of ensuring that nothing is missed while reviewing the case file.
Once the Arbitrator has cross-checked all the results of AI’s case file review, the Arbitrator will confirm in writing that in his/her discretion, the entire case file has been reviewed thoroughly and any AI mistakes/hallucinations regarding both the relevance of the documents and the conclusions drawn have been rectified. To ensure optimization of the use of both HI and AI, we are assuming here that while concluding the case file review phase, the Arbitrator has put all the critical information into context and understood and interpreted it with his/her intuition, compassion, cultural understanding and a sense of justice. The next phase of the adjudication process involves framing the issues. How much should the Arbitrator rely on AI to frame the issues?
* Rana Sajjad, a Columbia Law School alum, is a dual-qualified lawyer licensed as an Advocate of the Supreme Court of Pakistan and a Member of the New York Bar. He has over 25 years’ experience of practicing law in Pakistan and the U.S. in practice areas including contracts, cross-border transactions, commercial litigation and domestic and international arbitration. He is the Managing Partner at Triage Law, a Lahore-based commercial and arbitration law firm, and the Founder & President of the Center for International Investment and Commercial Arbitration (CIICA), Pakistan’s first international arbitration center.