Skip to content
Grounds publishedDecided 29 Sep 2025

High Court

Tajudin bin Gulam Rasul and another v Suriaya bte Haja Mohideen

Feedback & support

[2025] SGHCR 33Case no. Originating Claim No 125 of 2025 (Summons No 1240 of 2025)76 paragraphs

What happened, in plain English

There is no plain-English summary of this judgment yet.

Ask about this judgment

Laws this judgment relies on

Grounds of decision

Official text from eLitigation

  1. Introduction

  2. [1] Every advocate and solicitor of Singapore is an officer of the court, entrusted with the solemn duty of assisting the court in the administration of justice in Singapore (Law Society of Singapore v Udeh Kumar s/o Sethuraju and another matter [2017] 4 SLR 1369 at [104]). As members of a noble and honourable profession, advocates and solicitors are held to the highest standards of conduct. These standards are not mere aspirations – they comprise legally binding obligations put in place to safeguard the core values of the legal profession and preserve public trust and confidence in the legal profession and the administration of justice in Singapore.

  3. [2] Among these obligations lie the fundamental duty to ensure that all materials placed before the court, including any content generated by artificial intelligence (“AI”) tools, are independently verified, accurate, true and appropriate. It follows that the citation of a fictitious AI-generated authority to the court, whether intentionally or otherwise, is wholly impermissible.

  4. [3] When an advocate and solicitor cites a fictitious authority to the court, the gravity of his improper conduct does not lie solely in the loss of valuable judicial time and the unnecessary expenditure of his counterparty’s resources in uncovering his actions. Even more pernicious is the fissure that he foments in the public’s perception of the legal profession. His actions may, in the eyes of the public, cast a shadow over the legitimacy and honour of the legal profession and its role as a custodian of justice in Singapore.

  5. [4] The court will not condone such improper conduct.

  6. Factual background

  7. [6] I begin with an overview of the background facts.

  8. [7] SUM 1240 was an application (“Application”) by the Defendant in HC/OC 125/2025 (“OC 125”) to set aside a default judgment, HC/JUD 143/2025, which had been entered against her for failing to file a Notice of Intention to Contest or Not Contest the Claim. The central issue in SUM 1240 was whether the Defendant had a prima facie defence to the Claimants’ claim in OC 125.

  9. CC’s citation of a fictitious authority

  10. [8] Anticipating that the Defendant might raise a defence of illegality, the Claimants contended in their Written Submissions filed on 1 June 2025 (“CWS”) that any “defence under the Moneylenders Act [would be] inapplicable”. The Claimants cited a purported case and averred that it stood for the proposition that isolated or ad hoc loans between acquaintances do not constitute “moneylending” under the Moneylenders Act 2008 (2020 Rev Ed) unless there is a systematic business of lending for profit.

  11. [9] However, this case (“Fictitious Authority”) was conspicuously missing from the Claimants’ Bundle of Authorities filed on 2 June 2025 (“CBOA”).

  12. [10] It transpired that the case was fictitious.

  13. [11] In line with the practice adopted by the Federal Court of Australia in Luck v Secretary, Services Australia [2025] FCAFC 26 at [14], I have omitted the name and case number of the Fictitious Authority from these Grounds of Decision, so that false information will not be propagated further.

  14. Events leading up to the substantive hearing of SUM 1240

  15. [12] It is apposite to set out the timeline of material events leading up to the substantive hearing of SUM 1240 on 22 July 2025.

  16. [13] On 4 June 2025, I directed both parties to ensure that all cited authorities were exhibited and to insert pinpoint citations for all case references. The parties were informed of this direction on 5 June 2025.

  17. [14] According to the parties, on 18 July 2025, DC informed CC by way of email that he was unable to locate the Fictitious Authority cited in the CWS.

  18. [15] On the afternoon of 21 July 2025, CC filed amended written submissions (“AWS”) and a supplementary bundle of authorities containing a replacement authority (“CSBOA”) on e-Litigation. These documents were filed on CC’s own initiative, without seeking permission from the Court. Hard copies were tendered to the Court shortly thereafter.

  19. [16] I directed CC to identify the specific amendments made in the AWS and to provide reasons for these amendments. CC filed a letter on e-Litigation that afternoon, explaining that:

  20. [17] The hard copies of the AWS and CSBOA, and a cover letter dated 21 July 2025, were subsequently placed before me. The material parts of the cover letter are reproduced below:

  21. Hearing on 22 July 2025

  22. [18] During the hearing on 22 July 2025, CC informed me that he had filed the AWS the day before the hearing because of “issues with [paragraph] numbering” and because DC had pointed out that there was a case that was “wrongly cited”.

  23. The Defendant’s position

  24. [19] DC did not object to the Claimants’ filing of the AWS. He accepted that CC might have unintentionally cited the Fictitious Authority in the CWS.

  25. [20] However, DC submitted that CC was downplaying the severity of the matter. DC emphasised that he could not locate the non-existent case and that such an error was distinct from a mere typographical error like a misspelt party name.

  26. The Claimants’ position

  27. [21] CC accepted that DC’s allegation was “serious”.

  28. [22] However, he maintained that although the “citation [was] incorrect”, he had “not changed anything except [the] case [that] was wrong”.

  29. [23] Only when probed by the Court did CC finally admit that the Fictitious Authority “did not exist, which was why it was not [included] in the [CBOA]”. His explanation is set out below:

  30. The Court’s decision to permit the Claimants to tender the AWS and rely on the replacement authority

  31. [24] I noted the gravity of CC’s improper conduct and observed that it was CC’s responsibility to verify the existence of all authorities cited to the court.

  32. [25] Nevertheless, in circumstances where the Defendant did not object to the filing of the AWS and CSBOA, and where the legal proposition relied on by the Claimants was trite, I permitted the Claimants to tender the AWS in place of the CWS and rely on the replacement authority exhibited in the CSBOA.

  33. Defendant’s decision not to pursue the defence of illegality

  34. [26] Subsequently, in the course of the hearing, DC informed the Court that the Defendant would no longer pursue the defence of illegality but would continue to rely on other defences.

  35. Decision on merits

  36. [27] I eventually dismissed the Defendant’s Application on 31 July 2025, after considering the parties’ respective affidavits and submissions.

  37. Issue

  38. The parties’ submissions

  39. The Defendant’s position

  40. The Claimants’ position

  41. [32] CC advanced three main arguments in response:

  42. (a) First, although the Fictitious Authority did not exist, the legal proposition that CC sought to rely on existed. It was “in that circumstance” that CC had used a “placeholder case”;

  43. (b) Second, after “identif[ying] the error”, CC remedied it “as soon as [he] could” and “update[d] the Court” accordingly; and

  44. (c) Third, it was strange that DC ultimately did not rely on the defence of illegality and chose instead to withdraw it.

  45. [33] CC however acknowledged that there was an “error in the citation” that “should not have been there” and apologised “if the Court felt that [CC had] misled [it]”.

  46. [34] I afforded CC a further opportunity to be heard. When asked why a “placeholder” case was used and whether a GenAI tool was employed, CC admitted for the first time that:

  47. [35] CC identified the junior lawyer at the Court’s request, and stated as follows:

  48. Stance on the use of GenAI tools

  49. [36] CC’s citation of a fictitious AI-generated authority to the Court in the present case is lamentably symptomatic of an unsettling global trend of court users placing undue reliance on GenAI tools in preparing court documents, without verifying the accuracy of the AI-generated output. This trend has been observed across many jurisdictions, including Australia, Canada, New Zealand, the United Kingdom and the United States (Ayinde v London Borough of Haringey [2025] EWHC 1383 (Admin) (“Ayinde (Div Court)”) at [83]-[102]).

  50. [37] In Singapore, the use of GenAI tools in the preparation of documents for court proceedings is governed by the Guide on the Use of Generative Artificial Intelligence Tools by Court Users (“GenAI Guide”). The GenAI Guide applies to all matters in the Supreme Court (Supreme Court Registrar’s Circular No. 1 of 2024), as well as the State Courts (State Courts’ Registrar’s Circular No. 9 of 2024) and the Family Justice Courts (Family Justice Courts’ Registrar’s Circular No. 1 of 2024) (GenAI Guide at para 1).

  51. [38] The GenAI Guide provides a basic overview of how GenAI tools function. It explains that language models “predict” words to generate text, are unable to discern facts and are not designed to function as search engines (GenAI Guide at paras 4(2) and 4(3)). More importantly, it cautions that while outputs produced by GenAI tools may sound “persuasive and authoritative”, there is a risk that the output may be inaccurate or fabricated (ie, “hallucinated”) (GenAI Guide at para 4(3)).

  52. [39] The relevant provisions of the GenAI Guide are set out below:

  53. [40] As can be seen from the above, three main principles are embodied in the GenAI Guide:

  54. (a) First, the court maintains a neutral stance on the use of GenAI tools in the preparation of documents for court proceedings. This means that court users (including advocates and solicitors and self-represented persons) are not prohibited from using GenAI tools to prepare court documents, provided that they comply with the GenAI Guide (GenAI Guide at para 3(1)).

  55. (b) Second, court users must independently verify that all materials, including any AI-generated references to authorities, that are placed before the court exist and are accurate (GenAI Guide at paras 3(2), 5(1) and 5(3)). Advocates and solicitors bear an additional professional duty to do so (GenAI Guide at para 3(2)(a)).

  56. (c) Third, court users remain fully responsible for the content in all their court documents, and they must continue to comply with existing rules and practice directions, including the requirement to produce all cited authorities (GenAI Guide at paras 3(2), 3(3), 5(1) and 5(2)).

  57. [41] With these guardrails, the GenAI Guide seeks to support ethical, professional and responsible AI use so that the benefits of AI (such as enhancing efficiency for law practices and facilitating access to justice for self-represented persons) may be harnessed appropriately, without stifling innovation.

  58. The law

  59. Rationale behind imposing personal costs orders

  60. [45] The underlying principle behind this inherent power is that advocates and solicitors are officers of the court (section 82(1) of the Legal Profession Act 1966 (2020 Rev Ed) (“LPA”)) and the court has a right and a duty to supervise the conduct of its advocates and solicitors and penalise conduct that tends to defeat justice (Zhou Tong at [23]).

  61. [46] This is premised on two “practical and ethical considerations” (Zhou Tong at [25], citing Tan King Hiang v United Engineers (Singapore) Pte Ltd [2005] 3 SLR(R) 529 (“Tan King Hiang”) at [15]):

  62. (a) First, while advocates and solicitors have immunity from claims for negligence by their clients in respect of their conduct and management of a case in court and the pre-trial work immediately connected with it, they nevertheless have a duty to exercise reasonable care and skill in conducting their clients’ affairs; and

  63. (b) Second, a litigant should not be financially prejudiced by the unjustifiable conduct of litigation by the counterparty or the counterparty’s solicitor.

  64. The applicable test

  65. (a) Whether the advocate and solicitor acted improperly, unreasonably or negligently (“Stage 1”);

  66. (b) If so, whether the advocate and solicitor’s conduct caused the counterparty to incur unnecessary costs (“Stage 2”); and

  67. (c) If so, whether it is in all the circumstances just to order the advocate and solicitor to compensate the counterparty for the whole or any part of the relevant costs (“Stage 3”).

  68. [48] I will examine each stage in turn, in the specific context of an advocate and solicitor’s citation of a fictitious AI-generated authority to the court.

  69. Stage 1: Improper, unreasonable or negligent conduct

  70. [49] The approach to be taken in interpreting each of the words “improperly”, “unreasonably” and “negligently” has been considered in Ridehalgh at 232-233 and cited with approval in Tang Liang Hong at [71]:

  71. [50] These three terms are not mutually exclusive (Tan King Hiang at [19], citing Ridehalgh at 233). For instance, in Ayinde v London Borough of Haringey [2025] EWHC 1040 (Admin) (“Ayinde (Admin Court)”) at [64]-[65], Justice Ritchie found that “[i]t is wholly improper to put fake cases in a pleading”, “unreasonable, when it was pointed out, to say that these fake cases were “minor citation errors” or… “[c]osmetic errors”” and “negligent [to use] AI[,] not check it, [and] put that text into [a] pleading”.

  72. [51] In my judgment, in determining the threshold question of whether an advocate and solicitor has acted improperly, unreasonably or negligently in the specific context of his citation of a fictitious AI-generated authority to the court, the court should consider the standard of conduct expected of advocates and solicitors as set out in the GenAI Guide. The court may also, where appropriate, take into account the relevant provisions in the SupCt PD and the Legal Profession (Professional Conduct) Rules 2015 (“LP(PC)R”).

  73. (1) GenAI Guide

  74. [52] The provisions of the GenAI Guide have been discussed at [38] to [40] above. In essence, the “human-in-the-loop” or “lawyer-in-the-loop” approach, which requires independent human verification of AI-generated output before it is placed before the court, must be adopted. While GenAI tools may serve as an aid in the legal research process, advocates and solicitors have a professional, non-delegable duty to ensure that all materials placed before the court exist and are accurate.

  75. (2) SupCt PD

  76. [53] This safeguard is not a novel one. It has always existed as a fundamental tenet of our legal system. The SupCt PD, which sets out the procedural directions that advocates and solicitors must comply with, contains detailed rules on the preparation and filing of bundles of authorities. Specifically, para 101(8)(a) (read with paras 105(1) and 105(6)) of the SupCt PD obliges advocates and solicitors to include all authorities relied on in their respective bundles of authorities:

  77. [54] It follows that advocates and solicitors must, minimally, exercise due diligence to ensure that all authorities cited exist and have been exhibited.

  78. [55] The GenAI Guide does not alter the duty of advocates and solicitors to comply with these practice directions (GenAI Guide at paras 3(2) and 5(2)). It instead complements them by requiring final human oversight of AI-generated outputs.

  79. (3) LP(PC)R

  80. [56] In addition, ethics is the golden thread of legal practice, and advocates and solicitors are bound by rules of professional conduct (GenAI Guide at para 3(2)(a)). The majority of these rules have been codified in the LP(PC)R, which underscores an advocate and solicitor’s paramount duty to the court that must at all times be fulfilled in a manner that upholds the standing and integrity of the Singapore legal system and the legal profession in Singapore (rule 4(b) of the LP(PC)R).

  81. [57] An advocate and solicitor’s relationship with his clients and his conduct in the course of court proceedings is governed by various ethical obligations, including his duties to act with reasonable diligence and competence, not to knowingly mislead or attempt to mislead the court when conducting court proceedings, not to advance any proposition that he knows or ought reasonably to know is contrary to the law, to disclose to the court any procedural irregularity of which he is aware and to exercise proper supervision over the staff working under him in the law practice (rules 5(2)(c), 9(2)(a), 9(2)(f), 9(3)(b)(ii) and 32 of the LP(PC)R).

  82. [59] It bears emphasis that given that this is not a disciplinary inquiry, nothing in these Grounds of Decision shall be construed as the Court’s view on whether CC has breached any professional conduct rule.

  83. Stage 2: Unnecessary costs

  84. Stage 3: Just compensation for costs

  85. [65] In my view, while this inquiry is fact-sensitive and entails careful consideration of the specific circumstances of each case, the court may have regard to three broad principles:

  86. (a) Fairness and proportionality;

  87. (b) The need to preserve the integrity of the justice system; and

  88. (c) Deterrence.

  89. [66] I will address each of these principles in turn.

  90. (1) Fairness and proportionality

  91. [68] In my judgment, this entails a principled assessment of the specific facts of the case, taking into account factors such as:

  92. (a) The nature and impact of the advocate and solicitor’s conduct;

  93. (b) The prejudice suffered by the counterparty; and

  94. (c) The egregiousness of the advocate and solicitor’s conduct.

  95. [70] In considering the nature and impact of the advocate and solicitor’s conduct, the court may weigh the probability that the citation of the fictitious AI-generated authority could have potentially affected the court’s reasoning and/or the outcome of the proceedings. A fictitious authority cited in support of a central legal argument on which the outcome of the case may turn would pose a higher risk of materially affecting the court proceedings, and would in turn weigh more heavily on the court’s assessment of the gravity of the advocate and solicitor’s conduct. Conversely, if the fictitious authority had been cited in support of a peripheral issue that would have no effect on the court’s legal analysis and/or findings, its citation, while still wrongful, may be less material.

  96. [71] In relation to the prejudice suffered by the counterparty, the court should consider the burden imposed on the counterparty, as described at [61] above. The greater the burden, the more serious the prejudice suffered by the counterparty would have been.

  97. [72] As to the egregiousness of the advocate and solicitor’s conduct, the court may consider, among other things:

  98. (a) Whether the fictitious AI-generated authority was intentionally cited to mislead or deceive the court (Zhang v Chen 2024 BCSC 285 (“Zhang”) at [31]; Byoplanet International, LLC v Peter Johansson, No 0:25-cv-60630-LEIBOWITZ 2025 WL 2091025, at 11 (SD Fla, July 17, 2025) (“Byoplanet”));

  99. (b) Whether the advocate and solicitor had previously cited fictitious AI-generated authorities to the court (Byoplanet at 11);

  100. (c) Whether an immediate, full and truthful explanation is given to the court and the counterparty (Ayinde (Div Court) at [24]). Specifically, whether the advocate and solicitor expeditiously informed the court that a fictitious AI-generated authority was cited and took appropriate steps, in consultation with the court, to remedy the mistake, or conversely, if he displayed a lack of candour and attempted to downplay or conceal his mistake (see Byoplanet at 11); and

  101. (d) The impact on the underlying litigation (Ayinde (Div Court) at [24]), in particular, whether the legal proposition purportedly supported by the fictitious authority exists and could have been supported by a genuine authority.

  102. [73] Although the cases cited above do not expressly set out the factors that a court may consider in determining the egregiousness of an advocate and solicitor’s conduct, they nevertheless remain relevant insofar as they shed light on the principles that may guide the court’s decision.

  103. (2) Preserving the integrity of the justice system

  104. [75] As explained at [67] above, this is a subsidiary consideration, since the court’s primary focus in the present case is to ensure fairness between the parties.

  105. [76] Nevertheless, it is important for the court to enforce proper standards of conduct. The court should send a clear message that it is impermissible for advocates and solicitors to cite fictitious authorities to the court. Such improper conduct strikes at the heart of the common law legal system, for which case law is its “lifeblood” (Lee Tat Development Pte Ltd v Management Corporation Strata Title Plan No 301 [2018] 2 SLR 866 at [68]). Under this system, in line with the principle of stare decisis, the parties cite case law to persuade the court to follow similar precedents and the court “adjudicate[s] with reference to decided cases” (See Toh Siew Kee v Ho Ah Lam Ferrocement (Pte) Ltd and others [2013] 3 SLR 284 at [35]).

  106. [77] As Chief Justice Sundaresh Menon observed extrajudicially, “truth is the foundation of [adjudication]” (Chief Justice Sundaresh Menon, Supreme Court of Singapore, “Judicial Responsibility in the Age of Artificial Intelligence”, keynote speech at the Inaugural Singapore-India Conference on Technology at para 14 (15 April 2024) (accessed 29 September 2025)). A decision grounded on a fictitious authority may result in an “outrageous miscarriage of justice” (Ko v Li 2025 ONSC 2965 at [16]) and would erode public trust and confidence in the justice system.

  107. (3) Deterrence

  108. [79] The Court may also consider the principle of deterrence, since the advocate and solicitor would have made the active choice to rely on a GenAI tool to prepare documents for court proceedings without verifying the accuracy of its output. The consequence imposed should be sufficiently severe to deter the individual advocate and solicitor from repeating his improper conduct and to send a strong signal to other advocates and solicitors that such improper conduct will not be tolerated. Otherwise, an unduly lenient consequence may send the wrong signal that the court views such conduct lightly.

  109. Decision

  110. [80] Applying the law to the facts of the present case, I answered all three questions of the Ridehalgh test in the affirmative.

  111. Stage 1: CC’s improper, unreasonable and negligent conduct

  112. [81] I find that CC’s conduct falls within the ambit of “improper”, “unreasonable” and “negligent” conduct under the Ridehalgh test.

  113. [82] CC’s conduct is improper.

  114. (a) First, CC cited a fictitious AI-generated authority in the CWS. A simple search on the legal research portal LawNet would have revealed that the Fictitious Authority does not exist – the case name was fabricated and the citation number, while genuine, pertained to a wholly unrelated matter. CC’s citation of the fictitious authority and failure to verify the accuracy of the AI-generated content constitutes a non-compliance with the GenAI Guide. It also amounts to an abuse of court process (Zhang at [29]).

  115. (b) Second, CC failed to exercise due diligence in preparing the CBOA. He did not ensure its completeness and accuracy, and failed to exhibit all cited authorities in the CBOA. In so doing, he has breached not only para 101(8)(a) read with paras 105(1) and 105(6) of the SupCt PD, but also the Court’s express direction that all cited authorities must be exhibited and pinpoint citations provided.

  116. (c) Third, and of gravest concern, CC was less than candid with the Court and sought to downplay the gravity of his improper conduct.

  117. (i) As explained at [14] to [15] above, after DC alerted CC on 18 July 2025 that the cited case in the CWS was non-existent, CC failed to expeditiously inform the Court of the same. Instead, CC filed the AWS and introduced a replacement authority via the CSBOA on 21 July 2025 without the Court’s permission. While this may appear to be CC’s prompt remedy of his citation of the Fictitious Authority, CC’s lack of candour in explaining his actions is instead indicative of his attempt to conceal his improper conduct. He initially characterised it as a mere “clerical” or “typographical” error and stated that the Fictitious Authority was “wrongly” or “inadvertently” cited (see [16] to [18] above). It was only upon the Court’s questioning at the hearing on 22 July 2025 that CC admitted that he had cited a non-existent case (see [23] above). There is a material difference between typographical or citation errors and non-existent cases (Steven E Hobbs. Sr. v. Igor Goncharko, et. al., No. 25 CV 3398 2025 WL 2532561 at fn 7 (E.D.Ill. September 3 2025)). CC erred in conflating the two. I therefore categorically reject CC’s submission that he had remedied the error “as soon as [he] could” and thereafter “update[d] the Court”.

  118. (ii) CC also failed to voluntarily disclose that the Fictitious Authority was generated by a GenAI tool. This was only revealed upon the Court’s questioning during costs submissions on 31 July 2025 (see [34] above).

  119. (iii) In addition, CC described the Fictitious Authority as a “placeholder case” in the CWS, despite knowing that it was a fictitious AI-generated authority (see [34] above). The use of the word “placeholder” connotes an inadvertent inclusion of a temporary reference in the CWS. In describing the Fictitious Authority as a “placeholder case”, CC trivialised the seriousness of his improper conduct and demonstrated a continued failure to appreciate its gravity.

  120. (iv) Lastly, viewed in light of the fact that the sole issue that parties were asked to address before the Court was CC’s citation of the Fictitious Authority, CC’s explanation that the errors in the paragraph numbering in the CWS were attributable to his secretary (see [35] above), was an attempt to shy away from the central issue.

  121. [83] CC’s failure to comply with the GenAI Guide and SupCt PD, his lack of candour with the Court and his attempt to downplay the gravity of his conduct would be regarded as “improper”. As Justice Aidan Xu observed extrajudicially, advocates and solicitors “should know better” than to tender fictitious case citations (Justice Aidan Xu, Supreme Court of Singapore, “Legal and Regulatory Issues with Artificial Intelligence”, speech at the IT Law Series 2025 at para 3 (6 August 2025) (accessed 29 September 2025)).

  122. [84] CC’s conduct is also unreasonable. His citation of the Fictitious Authority permits of no reasonable explanation. As the United States District Court of the Southern District of Florida stated in Byoplanet at 12, a reasonable advocate and solicitor “does not blindly rely on AI [tools] to generate [court documents]”. His subsequent attempts to characterise it as a mere “clerical” and “typographical” error, as having been “wrongly” or “inadvertently” cited, or as serving merely as a “placeholder” are equally unjustifiable (see, for instance, Ayinde (Admin Court) at [64]).

  123. [85] CC’s conduct also constitutes negligent conduct under Stage 1 of the Ridehalgh test (Ayinde (Admin Court) at [65]), as he has failed to act with the competence reasonably expected of advocates and solicitors. His want of care in his conduct of the court proceedings and abdication of his professional duties is apparent from:

  124. (a) His failure to verify the accuracy of the AI-generated output and his consequent citation of a fictitious AI-generated authority to the Court (see [82(a)] above); and

  125. (b) His failure to ensure that all authorities have been exhibited in the CBOA, despite the provisions of the SupCt PD and the Court’s express direction (see [82(b)] above). Had he taken steps to comply with these directions, he would have realised that the Fictitious Authority does not exist.

  126. [86] Stage 1 of the Ridehalgh test has thus been fulfilled.

  127. Stage 2: Unnecessary costs incurred by the Defendant

  128. [88] Stage 2 of the Ridehalgh test is satisfied.

  129. Stage 3: Just compensation for the Defendant’s costs

  130. Fairness and proportionality

  131. [90] I begin with an examination of the principle of fairness and proportionality.

  132. [91] I accept that CC’s citation of the Fictitious Authority did not materially affect the Court’s reasoning or the outcome of these proceedings because (a) the Fictitious Authority was cited for a trite proposition and (b) DC eventually elected not to pursue the line of argument for which the Fictitious Authority was cited.

  133. [92] Nevertheless, this must be weighed against two other factors, viz, the prejudice suffered by the counterparty (as explained at [87] above) and the egregiousness of CC’s conduct (as explained at [82] to [85] above). In relation to the latter, I was mindful that while CC conceded that he “should have checked” the CWS and “ensure[d] [its] correctness” (see [34] to [35] above), his explanations remained troubling. CC’s statement that he would “take full responsibility for [the] error” because his colleague was newly called to the Bar and his apology to the Court if the Court “felt… misled” (see [33] above) demonstrated his failure to fully appreciate the gravity of his improper conduct.

  134. [93] For completeness, I gave no weight to CC’s attribution of the fictitious citation to the error of the alleged junior lawyer (see [23] and [34] above). Apart from the fact that there was no evidence of the same adduced before me, it must also be underscored that CC, the senior lawyer on the file, had a duty to supervise the work of his junior colleague. Furthermore, as the sole claimants’ counsel who appeared before me to make submissions for all hearings in respect of SUM 1240, he was expected to have read, understood and properly assessed the relevance of each case before citing them to the Court and informing me during the hearing on 4 June 2025 (after the CWS was filed) that he had “prepare[d] for [the] hearing” and was ready to proceed.

  135. Preserving the integrity of the justice system

  136. Deterrence

  137. [96] Stage 3 of the Ridehalgh test is fulfilled.

  138. Quantum of personal costs

  139. Orders made

  140. [100] In addition, given the severity of this matter, I further directed that both CC and DC shall provide their respective clients with a copy of my directions, to assure the parties that the court takes a serious view of such improper conduct by advocates and solicitors.

  141. Conclusion

  142. [101] In closing, advocates and solicitors must recognise that while technological tools may enhance the practice of law, these tools also have their inherent limitations. This incident serves as a solemn reminder that every advocate and solicitor bears a personal responsibility to comply with his or her professional duties. It is impermissible for an advocate and solicitor to delegate tasks such as conducting legal research and drafting written submissions to GenAI tools, without independently verifying the AI-generated output.

  143. [102] The practice of law must always remain firmly rooted in the core values of integrity, professionalism and justice: Singapore (Singapore), Final Report of the Ethics and Professional Standards Committee, 8 January 2025 at Annex C para 2. The public’s trust in the legal profession and the administration of justice in Singapore rests on the unwavering compliance of every advocate and solicitor with their expected standards of conduct, including those enshrined in the GenAI Guide. This trust should not be taken lightly and must never be broken.