A comparative study of judicial responses to AI use in legal practice, highlighting ethical limits, lawyer accountability, and emerging standards across leading common law jurisdictions—with insights from the ADGM Courts, the Indian judiciary, and the UK Courts on professional responsibility and ethical safeguards.
Introduction
The growth of Artificial Intelligence (AI) has been rapid, and it has been widely adopted as a tool to facilitate every aspect of our lives. AI has offered increased efficiency and productivity levels, saved time, lowered costs, and reduced human error. The legal field has adopted AI to revolutionise the complex system by automating tasks like legal research, document review (e-discovery), contract analysis, and case management, improving efficiency, speed, and accuracy for lawyers, judges, and courts.
However, AI is not foolproof. It is not uncommon to hear “AI is going to replace us”. While it provides for numerous new-technology-based benefits, it also presents several disadvantages and challenges.
A few AI disadvantages include:
- Mass job displacement for people whose work can be more efficiently done with AI (eg; data entry),
- Possible bias and discrimination from skewed data that the AI is trained on
- Major privacy/ security risks
- Decline in human creativity/ empathy from potential over-reliance, leading to skill decline
- Misinformation
- Environmental harm from intensive training.
Judicial Scrutiny of AI Use in Legal Practice: The ADGM Perspective
Al is transforming the legal system. From contract analysis and legal drafting to predicting case outcomes and assisting illiterate prisoners, to speeding up research and improving access to justice via chatbots, AI has done a lot for the field. However, AI cannot be used as a replacement for human function. Ethical use of AI is advocated for, and to uphold that, robust governance must be introduced. The judiciary in the UAE has emphasised the wise use of AI.
Arabyads Holding Limited v Gulrez Alam Marghoob Alam ADGMCFI 0032
The case of Arabyads Holding Limited v Gulrez Alam Marghoob Alam ADGMCFI 0032 is a significant judgment from the Abu Dhabi Global Market (ADGM) Courts regarding the professional obligations of lawyers when using artificial intelligence (AI). Justice Paul Heath KC issued this costs judgment primarily to address the circumstances in which legal representatives should be held liable for wasted costs due to the inappropriate use of AI in court documents.
Factual Background of the Dispute
The dispute originated when Arabyads alleged that Mr. Alam breached a Grant Agreement and an Employee Share Option Plan (ESOP). The company sought monetary damages and declarations that Mr. Alam was a “Bad Leaver,” which would allow them to reclaim certain shares. The total equity value of the claim was estimated to exceed USD 11.2 million. Represented by MIO Legal Consultants LLP (MIO), Mr. Alam filed a Defence on June 3, 2025. This document was described by the Court as “extraordinarily lengthy,” with the narrative alone spanning 47 pages and the total filing, including exhibits, reaching 233 pages. Arabyads challenged the Defence, asserting it contained numerous false legal authorities and was prepared using AI tools in a manner that violated professional standards. Specifically, the filing included fictitious cases such as “Allied Dunbar “ and “Johnston v Moreton “. It also mis-cited real cases, such as Caparo Industries Plc v Dickman, to support propositions they did not actually stand for. On the eve of a contested hearing in July 2025, Mr. Alam ended MIO’s representation, decided to act in person, and subsequently withdrew several pending applications.
Submissions of the Parties
The Claimant, Arabyads, claimed wasted costs for the improper use of AI which they contended was reckless and contrary to their duty to the court. They also sought ot have all their costs assessed on an indemnity basis to claim the highest level of cost recover against both MIO and Mr Alam.
The Defendant, Mr Alam, contended that that MIO was solely responsible for the defects in the Defence and should be liable for any wasted costs and sought to recover fees he had already paid to MIO. He further argued that the costs for his withdrawn applications should be assessed on the standard basis rather than the indemnity basis.
MIO argued they were under “severe practical constraints” because Mr. Alam had failed to pay for the legal work and refused to engage UK-qualified counsel to assist with the complex research. They further claimed AI was used only for formatting and structuring, not research, and that any errors were unintentional oversights caused by tight deadlines and lack of resources.
Findings and Holdings of the ADGM Court
The Court ordered the Defendant’s former legal firm, MIO Legal Consultants LLP (MIO), to pay the Claimant AED 282,508 in wasted costs. These costs were assessed on an indemnity basis, the highest possible level of cost recovery, because the Court found the firm’s conduct to be “unreasonable, if not improper.” The Court determined that MIO’s failure to verify AI-generated research—which resulted in the citation of non-existent “hallucinated” cases and the misapplication of real authorities and that this amounted to reckless conduct and a breach of the ADGM Court Rules of Conduct 2016.
Regarding the Transfer, Joinder, and Counterclaim applications, the Court held that the Defendant, Mr. Alam, must pay the Claimant AED 245,000. Unlike the wasted costs for the AI-related issues, these costs were assessed on a standard basis, meaning they had to be proportionate to the matters in issue.
The court, in its judgment stressed the importance of recognising the Lawyer’s Duty. It emphasises that while AI can be a useful tool, it does not absolve a lawyer of their fundamental duty to the Court. Justice Heath noted that the responsibility for “hallucinations” lies with the researcher, not the software, emphasizing the necessity of human oversight.
The court puts the onus on a legal practitioner using AI for research purposes to verify the existence of authorities on which they wish to rely and to confirm that they stand for the propositions for which they are being offered for opposing counsel’s consideration, and the Court. Without undertaking that verification task, the lawyer runs a serious risk that the Court may be misled. Justice Heath KC stated:
“In my view, lawyers using Al tools for research purposes should start from the premise that all authorities and/or articles on a particular topic that are revealed by Al research may not necessarily be accurately summarised in the response, or indeed may not exist. That puts the onus on a legal practitioner using Al for research purposes to verify the existence of authorities on which they wish to rely and to confirm that they stand for the propositions for which they are being offered for opposing counsel’s consideration, and the Court. Without undertaking that verification task, the lawyer runs a serious risk that the Court may be misled.”
Approaches Adopted in Other Jurisdictions
The Indian Judiciary’s Approach to Artificial Intelligence
In a keynote address at the Indo-Singapore Judicial Conference, in an event focused on ethical technology integration, cross-cultural legal exchange, and the importance of AI in legal proceedings, Chief Justice of India DY Chandrachud emphasised the ethical use of AI. While emphasising the crucial role of technology, particularly AI, in reshaping legal research and the judiciary, he also stressed the importance of ethical considerations in its integration in the profession.
He lauded Singapore for its emergence as a global technology and innovation Hub and acknowledged India’s strides in employing AI in the profession through new technology, such as the e-Courts project, which computerises court processes, digitises case records, and establishes online case management systems across all levels of the judiciary, the live transcription devices used in court which have increased access to justice through translations in several different languages and even the notable use of ChatGPT being utilised in adjudication.
Chandrachud believes the use of AI in the modern legal arena is inevitable, but it must come with “nuanced deliberation”. He noted the importance of transparency, accountability, and fairness in AI utilisation, highlighting concerns about potential errors and biases inherent in AI systems.
The Supreme Court of India released White Paper on AI and the Judiciary through its Centre for Research and Planning (CRP) in November 2025, clearly asserting that AI should enhance—not replace—human decision-making. The primacy of judicial reasoning, discretion, and accountability remains non-negotiable, and AI must be used as a tool in addressing structural challenges that have long burdened the Indian judiciary, such as the pendency of cases, workload on judges and complicated research.
The United Kingdom’s Judicial Position on AI
In October 2025, The Courts and Tribunals Judiciary published an updated guidance to assist Judicial Office Holders in relation to the use of Artificial Intelligence (AI). The guidance for judicial office holders emphasizes that any use of AI must protect the integrity of the administration of justice, placing personal responsibility on the judge for any material produced in their name. Judicial office holders must understand that AI chatbots generate text based on statistical probability rather than authoritative facts, making them unsuitable for conducting new legal research or complex analysis but potentially useful for administrative tasks or summarizing large bodies of text. To maintain confidentiality and privacy, users should never input non-public or sensitive information into public AI tools, as these systems store and reuse data to respond to future queries. Additionally, users must remain vigilant against AI “hallucinations”—such as fictitious cases or laws—and inherent data biases, requiring rigorous independent verification of all AI-generated output. Finally, the judiciary must be alert to the use of AI by litigants and legal representatives, ensuring that all submissions are accurate and free from deceptive elements like deepfakes or hidden “white text”.
Lord Justice Birss, Lead Judge for Artificial Intelligence, said:
“The use of AI by the judiciary must be consistent with its overarching obligation to protect the integrity of the administration of justice and uphold the rule of law.”
Moving Forward: Artificial Intelligence as an Aid, Not an Arbiter of Justice
The stance is that AI must be used responsibly. Lawyers have to ensure that they develop a sufficient understanding of how their AI tools work and that it is to be used as a tool to assist, not replace, human judgment. Lawyers remain responsible for the work product, decisions, and ethical obligations of their practice. AI can produce errors, “hallucinations” (fabrications), or biased results and therefore AI-generated content, research, and analysis must be thoroughly reviewed and verified by a qualified legal professional before it is used or presented as final work, as it might otherwise lead to unfair, incorrect or discriminatory outcomes for clients. The use of technology is inevitable, but one must remember that AI remains just a tool.
