Analysis

The Artificial-Intelligence Bill Comes Due: Law-Firm Spending and the Work Lawyers Must Defend

Law firm artificial intelligence spending now reaches research, drafting, review, billing, and supervision. Lawyers must still verify, price, sign, and defend the work.

Daniel R. Neuhaus

Daniel R. Neuhaus

— 21 min read
Share
The Artificial-Intelligence Bill Comes Due: Law-Firm Spending and the Work Lawyers Must Defend

The legal question is whether lawyers can command the systems that now help produce the work they must verify, price, and defend.[1]

Introduction

Lawyers who practiced through online research, email, electronic filing, remote proceedings, and discovery platforms have earned their skepticism toward legal-technology forecasts.[2] Those tools became routine parts of practice, altering speed, volume, location, and convenience while leaving the lawyer’s duties of competence, confidentiality, candor, supervision, and reasonable fees in place.[3]

Electronic discovery remains an important exception because it has changed litigation procedures, evidence management, vendor dependence, staffing, sanctions, and judicial supervision.[4] It supplies the control case for this article because electronically stored information forced lawyers and courts to build a rule-bound system for preserving, collecting, reviewing, producing, and defending digital evidence.[5]

Artificial intelligence has yet to produce an equivalent transformation of federal litigation procedure, although it now reaches the legal-production system through which firms research, draft, summarize, review, supervise, price, and deliver work.[6] The legal question has shifted from whether a machine may assist a lawyer to whether the lawyer can preserve professional control over the work that the machine helps produce.[7]

That question makes the current spending cycle a professional responsibility and client value problem.[8] Artificial intelligence can make good lawyers even more valuable when they use it to reduce waste, improve source discipline, sharpen budgets, and free up more time for judgment.[9] The same tools can let weak lawyers hide laziness behind fluent output, producing work that appears finished before anyone has tested whether it is true.[10]

Electronic discovery as the control case

Electronic discovery became the last major court-rule-bound legal-technology shift because electronically stored information changed the legal treatment of evidence.[11] A litigation file increasingly meant more than correspondence folders, banker’s boxes, and paper records, as clients held relevant material in email servers, backup tapes, metadata fields, shared drives, mobile devices, databases, and retention systems.[12]

The Federal Rules of Civil Procedure eventually reflected that practical change.[13] Rule 34 now allows requests for documents or electronically stored information, including data compilations stored in any medium from which information can be obtained, and the 2006 Advisory Committee notes state that electronically stored information stands on equal footing with paper documents.[14]

Rule 37(e) gives the shift its clearest procedural form.[15] The rule applies when electronically stored information that should have been preserved in anticipation of or in the conduct of litigation is lost because reasonable steps were not taken and the information cannot be restored or replaced through additional discovery.[16] It then separates curative measures based on prejudice from the most severe sanctions, which require an intent to deprive another party of the information’s use in litigation.[17]

The 2015 Advisory Committee notes explain why the rule mattered.[18] Growing volumes of electronically stored information had produced serious preservation problems, divergent sanctions standards, and excessive preservation costs, while courts needed a framework that recognized the common-law preservation duty and the practical impossibility of perfect preservation.[19]

Electronic discovery supplies a legal comparator because it shows how technology becomes part of lawyering when courts, clients, rules, vendors, and professional duties converge around a workflow.[20] Lawyers had to understand client information systems, negotiate production requests, manage privilege review at scale, supervise vendors and review teams, defend search methods, and explain preservation decisions within a rule-bound sanctions structure.[21]

That history provides the present cycle with a legal measure.[22] Technology becomes important to lawyering when it changes the work lawyers must understand, the processes they must supervise, the costs clients must bear, and the account lawyers must give to courts, clients, and opposing parties.[23]

The present spending record matters because firms are treating technology as part of the operating system for legal work.[24] The 2026 Thomson Reuters Institute and Georgetown Law report describes a sharp increase in law-firm spending on technology and knowledge-management tools, with average spending growth of 9.7 percent and 10.5 percent, respectively.[25] The same report distinguishes generative artificial intelligence from earlier tools because it can draft briefs, analyze contracts, and synthesize case law in ways that may alter how legal work is done.[26]

Knowledge-management spending deserves special attention because it concerns a law firm’s institutional memory.[27] A firm that invests heavily in knowledge systems is investing in how prior research, deal documents, litigation work, matter histories, and lawyer expertise can be found and reused, which places the expenditure closer to the machinery of legal production than to a faster inbox or better calendar.[28]

Kirkland & Ellis supplies the public flagship example because Reuters reported in May 2026 that the firm planned to spend $500 million over three to four years developing a custom artificial-intelligence platform, beginning with $100 million in 2026.[29] The report also stated that the platform would draw on information from 250 Kirkland lawyers and involve more than 180 technology professionals inside and outside the firm.[30]

A half-billion-dollar internal platform is capital allocation inside an elite legal institution.[31]

Kirkland remains a poor proxy for the median firm, but the size and character of the investment show how far the issue has moved from individual lawyer convenience.[32] At least some leading firms now view artificial intelligence as infrastructure tied to lawyer work, firm memory, and client service.[33]

Vendor growth supplies a related, more limited signal because funding rounds, valuations, integrations, and user counts indicate that investors and platform companies expect legal workflow to become a valuable market, although those measures do not establish accuracy, pricing reform, client savings, or improved professional judgment.[34] The public record is strongest when those signals are read together, with firm spending showing institutional commitment, vendor capitalization showing market expectation, and client adoption showing pressure from the buyer side.[35]

Why artificial intelligence is broader

Breadth is the stronger claim because the American Bar Association Formal Opinion 512 begins with familiar uses of artificial intelligence in electronic discovery and contract analytics, then states that generative artificial intelligence tools may assist with legal research, contract review, due diligence, document review, regulatory compliance, and drafting legal documents.[36]

That range matters because the relevant tasks are those in which lawyers convert law, facts, and client materials into work that another person may rely on.[37] Earlier tools altered discrete stages of practice, as online research made sources easier to retrieve, electronic filing moved papers through court systems, and electronic discovery required litigators to manage digital evidence under preservation and production rules.[38] Generative artificial intelligence reaches closer to the lawyer’s own work product because it can help produce analysis, summaries, comparisons, drafts, and recommendations that may later be communicated, filed, signed, supervised, or billed.[39]

The professional problem travels across practice groups because the duties attach to the lawyer’s reliance on the output rather than to the subject matter of the task.[40] A lawyer who uses a tool to compare deal documents, summarize agency materials, or draft a litigation research section must still decide whether the output is reliable for that assignment.[41] That decision requires the lawyer to identify the supporting sources, account for omissions, and determine the professional consequence that follows from using the work.[42]

Knowledge management extends the problem because a firmwide artificial-intelligence system can make prior work, similar matters, client-document summaries, research memoranda, and internal forms available across the firm.[43] Careful governance can reduce duplication and improve consistency by allowing lawyers to locate the firm’s own work before rebuilding a research path, diligence process, or draft from the beginning.[44] Poor governance can spread stale law, recycle weak drafting, obscure the source and limits of prior work, expose client information, and give junior lawyers confidence in material they lack enough context to test.[45]

The point is broader than tool adoption because artificial intelligence now operates near the place where information becomes legal work.[46] When a system helps produce work that a lawyer may communicate, file, supervise, charge for, or defend, the governing question becomes professional responsibility.[47]

Professional responsibility supplies the rule

Professional responsibility gives the current cycle its legal center because the lawyer’s duties attach when the lawyer uses, communicates, supervises, files, or bills work shaped by artificial intelligence.[48] Formal Opinion 512 treats that use as governed by familiar duties of competence, client protection, communication, supervision, candor, meritorious advocacy, and reasonable fees.[49]

That approach is institutionally sound because the legal system can apply settled professional duties to concrete uses of a tool without first judging artificial intelligence in the abstract.[50] When the technology contributes to research, advice, a filing, or a bill, the lawyer remains answerable for the professional act that follows and for the judgment required to defend it.[51]

Competence now requires a working understanding of the particular tool being used, even though Formal Opinion 512 does not require lawyers to become generative-artificial-intelligence experts.[52] The opinion requires sufficient understanding to assess the tool’s capabilities and limitations, including its benefits, risks, and potential to produce inaccurate, incomplete, unreliable, or fabricated output.[53]

Formal Opinion 512 also states the nondelegation principle directly, allowing generative artificial intelligence to supply a draft or analytical starting point while preserving the lawyer’s responsibility for work that requires professional judgment.[54] Supervision gives that rule institutional form, since managerial lawyers must adopt effective compliance measures and supervisory lawyers must make reasonable efforts to ensure proper use by subordinate lawyers and nonlawyers.[55]

Artificial intelligence may change the means by which legal work is produced, although it does not change who must answer for the work.[56] The lawyer’s duties attach when the lawyer uses the tool, reviews the output, communicates the answer, signs the filing, bills the client, or defends the work after a challenge.[57]

Verification is the first application

Verification first tests lawyer judgment because artificial intelligence can make weak work appear complete, giving a research section or motion draft the surface order of legal analysis before the authorities have been checked.[58] A fluent passage may contain invented authority, distorted quotations, a false procedural posture, or a conclusion that the cited source cannot support, and the surrounding polish can make those defects harder to detect.[59]

The risk has empirical support because Formal Opinion 512 cites research finding that leading legal-research generative artificial intelligence systems hallucinated between 17 percent and 33 percent of the time in the tested study, which the opinion uses to support competence and verification duties.[60]

Mata v. Avianca remains the canonical district-court warning because Judge P. Kevin Castel, a Senior Judge of the United States District Court for the Southern District of New York, distinguished permissible assistance from professional abdication.[61] He acknowledged that lawyers may use reliable artificial-intelligence tools, yet sanctioned attorneys who submitted nonexistent opinions with false quotations and citations, then continued to defend those materials after the court questioned whether they existed.[62]

Mata matters because the lawyers presented false legal material before a federal court without checking it with ordinary legal sources, and their failure to correct the defect after warning signs appeared turned a research failure into a candor problem.[63]

The Ninth Circuit’s 2026 order in LNU v. Blanche strengthens the point by placing the violation at the professional act of signing and filing.[64] The panel imposed sanctions after briefs contained nonexistent cases, misattributed quotations, and serious misrepresentations of actual authorities, while explaining that the misconduct lay in filing false work rather than in the earlier use of generative artificial intelligence.[65]

Existing duties already reach that conduct because courts can address false authority through ordinary rules of candor, signature responsibility, and sanctions authority.[66] A lawyer violates those obligations by filing nonexistent cases, quoting language an opinion does not contain, or attributing an unsupported proposition to a real decision, regardless of whether artificial intelligence explains how the error entered the draft.[67]

New York’s statewide Part 161 follows the same premise from the rulemaking side.[68] The rule allows the use of artificial intelligence in preparing court papers when existing duties are satisfied, while its model provision still requires careful review and independent assurance that the paper contains no fabricated or fictitious legal material.[69]

Disclosure rules may differ by court, but the verification duty does not.[70] A court may require disclosure in a particular setting, and another court may rely solely on signature rules and candor duties, yet the lawyer must verify the work the lawyer files.[71]

Who receives the value of efficiency?

When artificial intelligence reduces the time required for research, review, drafting, diligence, or billing analysis, the gain must appear somewhere in the relationship between firm and client, and the resulting dispute is both ethical and commercial because candor, reasonable fees, and client trust all depend on whether the firm can explain where the efficiency went.[72]

The Thomson Reuters and Georgetown report provides a market context, as firms are buying tools that promise faster work while much of the legal market still sells lawyer time.[73] The report states that 90 percent of legal dollars still flow through standard hourly arrangements, as clients scrutinize line items and compare outside-counsel bills with efficiencies appearing in their own legal departments.[74]

Formal Opinion 512 supplies the ethical rule for that pricing pressure by allowing a lawyer to charge for the time actually spent using and reviewing generative-artificial-intelligence work while barring charges for time never spent or for time made necessary by the lawyer’s own inexperience with a tool the lawyer expects to use regularly.[75]

That rule preserves profit tied to defensible value rather than recorded fiction.[76] A client who buys time may be charged only for time worked, while a client who buys a fixed result may pay a reasonable fee measured by the agreement, the work required, the value delivered, and the professional responsibility the firm assumes.[77]

Client pressure gives the issue an institutional form because the Association of Corporate Counsel and Everlaw reported in October 2025 that 52 percent of surveyed in-house lawyers actively used generative artificial intelligence, more than double the 23 percent reported in 2024.[78] Many respondents also expected to rely less on law firms and handle more work internally.[79]

Those survey results are best read as evidence of market pressure because expectations alone do not alter engagement letters and adoption numbers do not establish durable savings.[80] The data still show why firms cannot treat artificial-intelligence use as a purely internal technology choice, since clients are beginning to ask whether efficiency appears in price, budget discipline, turnaround time, or improved work product.[81]

Used well, these tools can reduce waste, improve consistency, and allow lawyers to spend more time on judgment rather than retrieval, formatting, and first-pass synthesis.[82] The lawyer who commands the tool may deliver better work for better value, while the lawyer who defers to the tool may deliver fluent error at lower visible cost and higher professional risk.[83]

Governance connects the fee question to the client-value claim by giving a firm a way to preserve professional control over work produced through human and machine processes together.[84] A firm needs rules that identify approved tools, protect client information, require verification, assign supervision, address billing practices, and account for court-specific disclosure obligations when they apply.[85]

The narrower claim survives the counterargument

The strongest counterargument is that artificial intelligence may follow the familiar path of legal technology, moving from forecasted disruption into ordinary practice after lawyers, clients, and courts absorb its useful functions.[86] Electronic discovery remains the harder comparator on procedure because it produced a rule-based architecture for preservation, production, sanctions, proportionality, vendor management, and litigation support that artificial intelligence has yet to produce for ordinary litigation.[87]

Reliability evidence and market evidence both counsel restraint.[88] Formal Opinion 512 cites hallucination rates significant enough to require verification discipline, while Kirkland’s reported $500 million investment reflects the frontier rather than the median firm.[89] Many regional firms, small litigation shops, government offices, and solo practices lack the capital, technical staff, data discipline, training capacity, and client base needed to build or govern comparable systems.[90] Professional risk is hardest to control where governance capacity is weakest.[91]

Those objections narrow the thesis without defeating it.[92] Artificial intelligence has yet to displace electronic discovery as the leading example of technology remaking court-governed litigation practice, while its present claim rests on reaching more of the ordinary work lawyers perform before a court ever sees the result.[93]

Technology spending and the law of lawyering

Law-firm technology spending becomes legally significant once it reaches the processes by which lawyers produce client work, test its accuracy, supervise its preparation, price it, and deliver it.[94] The present cycle has reached that point because a firm’s ownership of an artificial-intelligence tool or pilot program means little unless the firm can account for the work the tool helps produce.[95]

That professional control has practical content because a firm must govern the tool before client information enters it, supervise the work while the tool is being used, and decide how the resulting work may be billed, disclosed, filed, or sent.[96] These are questions of competence, confidentiality, candor, supervision, and fees, even when they arise through software procurement or internal technology policy.[97]

Electronic discovery taught lawyers that technology becomes part of lawyering when legal duties, client expectations, vendor systems, court supervision, and cost converge around a workflow.[98] The present cycle now brings that lesson into the firm’s ordinary production of legal work, where the decisive question is whether the firm can command the system rather than merely own it.[99]

Artificial intelligence can improve legal service only when it remains subordinate to the duties that define lawyering.[100] The client hires a lawyer for judgment, and that judgment is measured by work the lawyer is willing to verify, supervise, price, sign, and defend.[101]

Notes

  1. See ABA Comm. on Ethics & Prof'l Responsibility, Formal Op. 512, Generative Artificial Intelligence Tools 1–6, 12–15 (2024); W. Bradley Wendel, The Promise and Limitations of Artificial Intelligence in the Practice of Law, 72 Okla. L. Rev. 21, 26–27, 40–43 (2019); LNU v. Blanche, No. 24-4790, slip op. at 2–5, 14–18 (9th Cir. June 3, 2026); Fed. R. Civ. P. 11(b)–(c). ↩︎

  2. See U.S. Courts, Electronic Filing (CM/ECF) (last visited June 18, 2026); PACER, Manage My Account Login (last visited June 18, 2026); U.S. Courts, Remote Public Access to Proceedings (last visited June 18, 2026); Thomson Reuters Inst. & Georgetown Law Ctr. on Ethics & the Legal Profession, 2026 Report on the State of the US Legal Market 11–12 (2026); Model Rules of Prof'l Conduct r. 1.1 cmt. 8 (Am. Bar Ass'n 2024). ↩︎

  3. See Model Rules of Prof'l Conduct rr. 1.1, 1.4, 1.5, 1.6, 3.1, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024); Formal Op. 512, supra note 1, at 1–6, 10–15. ↩︎

  4. See Fed. R. Civ. P. 34(a)(1)(A), (b)(1)(C), (b)(2)(D)–(E), 37(e); Fed. R. Civ. P. 34 advisory committee's note to 2006 amendment; Fed. R. Civ. P. 37 advisory committee's note to 2015 amendment; Ronald J. Hedges, Barbara J. Rothstein & Elizabeth C. Wiggins, Managing Discovery of Electronic Information 1–7, 31–35, 40–43 (Fed. Jud. Ctr. 3d ed. 2017, 2d printing 2019); The Sedona Conference, The Sedona Principles, Third Edition, 19 Sedona Conf. J. 1, principles 1–5 (2018); EDRM, EDRM Model (last visited June 18, 2026). ↩︎

  5. See sources cited supra note 4; Hon. Shira A. Scheindlin & Jeffrey Rabkin, Electronic Discovery in Federal Civil Litigation: Is Rule 34 Up to the Task?, 41 B.C. L. Rev. 327, 331–49 (2000); Gregory P. Joseph, Rule 37(e): The New Law of Electronic Spoliation, 99 Judicature No. 3 (2015). ↩︎

  6. See Formal Op. 512, supra note 1, at 1–6, 12–15; Ryan McCarl, The Limits of Law and AI, 90 U. Cin. L. Rev. 923, 923–49 (2022); Sayash Kapoor, Peter Henderson & Arvind Narayanan, Promises and Pitfalls of Artificial Intelligence for Legal Applications, 2 J. Cross-Disciplinary Rsch. Computational L. (2024); 22 N.Y.C.R.R. pt. 161, §§ 161.1–161.4 & app. A (effective June 1, 2026). ↩︎

  7. See Formal Op. 512, supra note 1, at 1–6, 12–15; Wendel, supra note 1, at 26–27, 40–43; McCarl, supra note 6, at 923–49; Kapoor, Henderson & Narayanan, supra note 6. ↩︎

  8. See Formal Op. 512, supra note 1, at 1–6, 10–15; Mata v. Avianca, Inc., 678 F. Supp. 3d 443, 448–49, 461–66 (S.D.N.Y. 2023); LNU, slip op. at 2–5, 14–25; Varun Magesh et al., Hallucination-Free? Assessing the Reliability of Leading AI Legal Research Tools, 22 J. Empirical Legal Stud. 216, 216–42 (2025); Fed. R. Civ. P. 11(b)–(c). ↩︎

  9. See Formal Op. 512, supra note 1, at 3–5, 10–12; Thomson Reuters Inst. & Georgetown Law Ctr. on Ethics & the Legal Profession, supra note 2, at 11–12. ↩︎

  10. See Mata, 678 F. Supp. 3d at 448–49, 461–66; LNU, slip op. at 2–5, 14–25; Magesh et al., supra note 8, at 216–42. ↩︎

  11. See sources cited supra notes 4–5. ↩︎

  12. See Fed. R. Civ. P. 34 advisory committee's note to 2006 amendment; Hedges, Rothstein & Wiggins, supra note 4, at 1–7, 31–35; Scheindlin & Rabkin, supra note 5, at 331–49; EDRM Model, supra note 4. ↩︎

  13. See Fed. R. Civ. P. 34(a)(1)(A), (b)(1)(C), (b)(2)(D)–(E); Fed. R. Civ. P. 34 advisory committee's note to 2006 amendment. ↩︎

  14. See Fed. R. Civ. P. 34(a)(1)(A), (b)(1)(C), (b)(2)(D)–(E); Fed. R. Civ. P. 34 advisory committee's note to 2006 amendment. ↩︎

  15. See Fed. R. Civ. P. 37(e); Fed. R. Civ. P. 37 advisory committee's note to 2015 amendment. ↩︎

  16. See Fed. R. Civ. P. 37(e). ↩︎

  17. See Fed. R. Civ. P. 37(e)(1)–(2). ↩︎

  18. See Fed. R. Civ. P. 37 advisory committee's note to 2015 amendment; Joseph, supra note 5. ↩︎

  19. See Fed. R. Civ. P. 37 advisory committee's note to 2015 amendment. ↩︎

  20. See Fed. R. Civ. P. 34(a)(1)(A), 37(e); Hedges, Rothstein & Wiggins, supra note 4, at 1–7, 31–35, 40–43; The Sedona Conference, supra note 4, principles 1–5; EDRM Model, supra note 4. ↩︎

  21. See Hedges, Rothstein & Wiggins, supra note 4, at 1–7, 31–35, 40–43; The Sedona Conference, supra note 4, principles 1–5; EDRM Model, supra note 4. ↩︎

  22. See sources cited supra notes 4–5, 8. ↩︎

  23. See sources cited supra notes 4–5, 8; Fed. R. Civ. P. 11(b)–(c). ↩︎

  24. See Thomson Reuters Inst. & Georgetown Law Ctr. on Ethics & the Legal Profession, supra note 2, at 11–12. ↩︎

  25. See id. at 11. ↩︎

  26. See id. ↩︎

  27. See Jon Beaumont, Knowledge Management: A Systems Case Study from Shearman & Sterling LLP, 17 Legal Info. Mgmt. 220, 220–28 (2017); Hélène Russell, A Law Firm Librarian's Guide to KM, 16 Legal Info. Mgmt. 131, 131–37 (2016); Petter Gottschalk & Vijay K. Khandelwal, Knowledge Management Technology in Law Firms: Stages of Growth, 18 Int'l Rev. L. Computers & Tech. 375, 375–85 (2004). ↩︎

  28. See Beaumont, supra note 27, at 220–28; Russell, supra note 27, at 131–37; Gottschalk & Khandelwal, supra note 27, at 375–85; Thomson Reuters Inst. & Georgetown Law Ctr. on Ethics & the Legal Profession, supra note 2, at 11. ↩︎

  29. See Mike Scarcella, Law Firm Kirkland to Spend $500 Million Developing Its Own AI Platform, Reuters (May 28, 2026). ↩︎

  30. See id. ↩︎

  31. See id. ↩︎

  32. See id.; Thomson Reuters Inst. & Georgetown Law Ctr. on Ethics & the Legal Profession, supra note 2, at 11–12. ↩︎

  33. See Scarcella, supra note 29; Thomson Reuters Inst. & Georgetown Law Ctr. on Ethics & the Legal Profession, supra note 2, at 11–12. ↩︎

  34. See Pragyan Kalita & Prakhar Srivastava, Legal Software Firm Harvey Valued at $11 Billion in Latest Funding Round, Reuters (Mar. 25, 2026); Prakhar Srivastava, Legal AI Firm Clio Valued at $5 Billion in Latest Funding Round, Reuters (Nov. 10, 2025); Thomson Reuters, One Million Professionals Turn to CoCounsel as Thomson Reuters Scales AI for Regulated Industries (Feb. 24, 2026); Ass'n of Corp. Couns. & Everlaw, New ACC Report Finds Generative AI Use in Corporate Law Departments More Than Doubled in a Single Year (Oct. 14, 2025). ↩︎

  35. See sources cited supra notes 24–34. ↩︎

  36. See Formal Op. 512, supra note 1, at 1. ↩︎

  37. See Formal Op. 512, supra note 1, at 1–6, 12–15; Wendel, supra note 1, at 26–27, 40–43; McCarl, supra note 6, at 923–49; Kapoor, Henderson & Narayanan, supra note 6; LNU, slip op. at 14–18. ↩︎

  38. See sources cited supra notes 2, 4–6. ↩︎

  39. See Formal Op. 512, supra note 1, at 1, 3–5; LNU, slip op. at 14–18; McCarl, supra note 6, at 923–49; Kapoor, Henderson & Narayanan, supra note 6. ↩︎

  40. See Model Rules of Prof'l Conduct rr. 1.1, 1.4, 1.5, 1.6, 3.1, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024); Formal Op. 512, supra note 1, at 1–15; LNU, slip op. at 14–18. ↩︎

  41. See Formal Op. 512, supra note 1, at 3–6; Model Rules of Prof'l Conduct r. 1.1 (Am. Bar Ass'n 2024). ↩︎

  42. See Formal Op. 512, supra note 1, at 3–6, 12–15; Model Rules of Prof'l Conduct rr. 1.1, 3.3 (Am. Bar Ass'n 2024); LNU, slip op. at 14–18. ↩︎

  43. See Formal Op. 512, supra note 1, at 1, 5–6; Beaumont, supra note 27, at 220–28; Russell, supra note 27, at 131–37; Gottschalk & Khandelwal, supra note 27, at 375–85. ↩︎

  44. See Beaumont, supra note 27, at 220–28; Russell, supra note 27, at 131–37; Gottschalk & Khandelwal, supra note 27, at 375–85. ↩︎

  45. See Formal Op. 512, supra note 1, at 3–6, 12–15; Beaumont, supra note 27, at 220–28; Russell, supra note 27, at 131–37; Gottschalk & Khandelwal, supra note 27, at 375–85. ↩︎

  46. See Model Rules of Prof'l Conduct rr. 1.1, 1.4, 1.5, 1.6, 3.1, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024); Formal Op. 512, supra note 1, at 1–15; Wendel, supra note 1, at 26–27, 40–43; LNU, slip op. at 14–18. ↩︎

  47. See Model Rules of Prof'l Conduct rr. 1.1, 1.4, 1.5, 1.6, 3.1, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024); Formal Op. 512, supra note 1, at 1–15. ↩︎

  48. See Model Rules of Prof'l Conduct rr. 1.1, 1.4, 1.5, 1.6, 3.1, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024); Formal Op. 512, supra note 1, at 1–15. ↩︎

  49. See Formal Op. 512, supra note 1, at 1. ↩︎

  50. See Model Rules of Prof'l Conduct rr. 1.1, 1.4, 1.5, 1.6, 3.1, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024); Formal Op. 512, supra note 1, at 1–15; Wendel, supra note 1, at 40–43. ↩︎

  51. See Formal Op. 512, supra note 1, at 3–6, 10–15; Model Rules of Prof'l Conduct rr. 1.1, 1.5, 3.1, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024). ↩︎

  52. See Formal Op. 512, supra note 1, at 2–4. ↩︎

  53. See id. at 3–4. ↩︎

  54. See id. at 4–5. ↩︎

  55. See id. at 7–10; Model Rules of Prof'l Conduct rr. 5.1, 5.3 (Am. Bar Ass'n 2024). ↩︎

  56. See Model Rules of Prof'l Conduct rr. 1.1, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024); Formal Op. 512, supra note 1, at 3–6, 7–10, 12–15; Mata, 678 F. Supp. 3d at 448–49, 461–66; LNU, slip op. at 14–18; Fed. R. Civ. P. 11(b)–(c). ↩︎

  57. See sources cited supra note 56. ↩︎

  58. See Formal Op. 512, supra note 1, at 3–5, 12–15; Mata, 678 F. Supp. 3d at 448–49, 461–66; LNU, slip op. at 14–18; Magesh et al., supra note 8, at 216–42. ↩︎

  59. See Mata, 678 F. Supp. 3d at 448–49, 461–66; LNU, slip op. at 14–18; Magesh et al., supra note 8, at 216–42. ↩︎

  60. See Formal Op. 512, supra note 1, at 3–4 n.14; Magesh et al., supra note 8, at 216–42. ↩︎

  61. See Mata, 678 F. Supp. 3d at 448–49, 461–66. ↩︎

  62. See id. at 448–49, 461–66. ↩︎

  63. See id.; Fed. R. Civ. P. 11(b)–(c). ↩︎

  64. See LNU, slip op. at 2–5, 14–18. ↩︎

  65. See id. ↩︎

  66. See Model Rules of Prof'l Conduct rr. 1.1, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024); Fed. R. Civ. P. 11(b)–(c); Mata, 678 F. Supp. 3d at 448–49, 461–66; LNU, slip op. at 14–18. ↩︎

  67. See sources cited supra note 66. ↩︎

  68. See 22 N.Y.C.R.R. pt. 161, §§ 161.1–161.4 & app. A (effective June 1, 2026). ↩︎

  69. See id. §§ 161.3–161.4 & app. A. ↩︎

  70. See Fed. R. Civ. P. 11(b)–(c); Mata, 678 F. Supp. 3d at 448–49, 461–66; LNU, slip op. at 14–18; 22 N.Y.C.R.R. pt. 161, §§ 161.3–161.4 & app. A. ↩︎

  71. See sources cited supra note 70. ↩︎

  72. See Model Rules of Prof'l Conduct rr. 1.1, 1.4, 1.5, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024); Formal Op. 512, supra note 1, at 10–12; Thomson Reuters Inst. & Georgetown Law Ctr. on Ethics & the Legal Profession, supra note 2, at 12; Ass'n of Corp. Couns. & Everlaw, supra note 34. ↩︎

  73. See Thomson Reuters Inst. & Georgetown Law Ctr. on Ethics & the Legal Profession, supra note 2, at 11–12. ↩︎

  74. See id. at 12. ↩︎

  75. See Formal Op. 512, supra note 1, at 10–12; Model Rules of Prof'l Conduct r. 1.5 (Am. Bar Ass'n 2024). ↩︎

  76. See Formal Op. 512, supra note 1, at 10–12; Model Rules of Prof'l Conduct r. 1.5 (Am. Bar Ass'n 2024). ↩︎

  77. See Formal Op. 512, supra note 1, at 10–12; Model Rules of Prof'l Conduct r. 1.5 (Am. Bar Ass'n 2024). ↩︎

  78. See Ass'n of Corp. Couns. & Everlaw, supra note 34. ↩︎

  79. See id. ↩︎

  80. See Thomson Reuters Inst. & Georgetown Law Ctr. on Ethics & the Legal Profession, supra note 2, at 12; Ass'n of Corp. Couns. & Everlaw, supra note 34. ↩︎

  81. See sources cited supra note 80. ↩︎

  82. See Formal Op. 512, supra note 1, at 3–5, 10–12; Thomson Reuters Inst. & Georgetown Law Ctr. on Ethics & the Legal Profession, supra note 2, at 11–12. ↩︎

  83. See Formal Op. 512, supra note 1, at 3–5, 12–15; Mata, 678 F. Supp. 3d at 448–49, 461–66; LNU, slip op. at 14–18; Magesh et al., supra note 8, at 216–42. ↩︎

  84. See Model Rules of Prof'l Conduct rr. 1.1, 1.4, 1.5, 1.6, 3.1, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024); Formal Op. 512, supra note 1, at 1–15; 22 N.Y.C.R.R. pt. 161, §§ 161.3–161.4 & app. A. ↩︎

  85. See sources cited supra note 84. ↩︎

  86. See Fed. R. Civ. P. 34(a)(1)(A), 37(e); Hedges, Rothstein & Wiggins, supra note 4, at 1–7, 31–35, 40–43; The Sedona Conference, supra note 4, principles 1–5; Wendel, supra note 1, at 21–43; McCarl, supra note 6, at 923–49; Kapoor, Henderson & Narayanan, supra note 6. ↩︎

  87. See sources cited supra notes 4–5, 86. ↩︎

  88. See Formal Op. 512, supra note 1, at 3–5; Scarcella, supra note 29; Magesh et al., supra note 8, at 216–42; LNU, slip op. at 14–18. ↩︎

  89. See Formal Op. 512, supra note 1, at 3–4 n.14; Magesh et al., supra note 8, at 216–42; Scarcella, supra note 29. ↩︎

  90. See Formal Op. 512, supra note 1, at 2–6, 7–12; Scarcella, supra note 29; Thomson Reuters Inst. & Georgetown Law Ctr. on Ethics & the Legal Profession, supra note 2, at 11–12. ↩︎

  91. See sources cited supra notes 88–90. ↩︎

  92. See sources cited supra notes 4–8, 86–91. ↩︎

  93. See sources cited supra notes 4–8, 36–47, 86–92. ↩︎

  94. See Model Rules of Prof'l Conduct rr. 1.1, 1.4, 1.5, 1.6, 3.1, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024); Formal Op. 512, supra note 1, at 1–15; Thomson Reuters Inst. & Georgetown Law Ctr. on Ethics & the Legal Profession, supra note 2, at 11–12; Scarcella, supra note 29; Ass'n of Corp. Couns. & Everlaw, supra note 34. ↩︎

  95. See sources cited supra note 94. ↩︎

  96. See Formal Op. 512, supra note 1, at 1–15; Model Rules of Prof'l Conduct rr. 1.1, 1.5, 1.6, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024); 22 N.Y.C.R.R. pt. 161, §§ 161.3–161.4 & app. A. ↩︎

  97. See sources cited supra note 96. ↩︎

  98. See Fed. R. Civ. P. 34(a)(1)(A), 37(e); Hedges, Rothstein & Wiggins, supra note 4, at 1–7, 31–35, 40–43; The Sedona Conference, supra note 4, principles 1–5; EDRM Model, supra note 4. ↩︎

  99. See sources cited supra notes 94–98. ↩︎

  100. See Model Rules of Prof'l Conduct rr. 1.1, 1.4, 1.5, 1.6, 3.1, 3.3, 5.1, 5.3 (Am. Bar Ass'n 2024); Formal Op. 512, supra note 1, at 1–15; Mata, 678 F. Supp. 3d at 448–49, 461–66; LNU, slip op. at 14–18; Fed. R. Civ. P. 11(b)–(c). ↩︎

  101. See sources cited supra note 100. ↩︎

Read more