The uncovered seat still has no AI interview scorecard.

A New York head of legal who has an uncovered Associate General Counsel or legal-operations seat is not short of licensed lawyers. The New York State Unified Court System’s year-end 2025 registration count put 193,536 attorneys in-state. Density is not a checklist. The skill the chief executive has started to ask for still has no written test behind it, and the person who has to invent that test is the same general counsel who already interviews for commercial, litigation, privacy and employment seats.

CLOC’s 2025 State of the Industry, based on the 2024 Harbor Law Department Survey of 186 organizations, is the map of where the tools actually sit. Fifty-seven percent of participating departments have a team reporting to a Director or Manager of Legal Operations; 22 percent have that director as the only formal operations resource; 11 percent have no formal resource; 10 percent decentralize operations. The director role is typically full-time (87 percent) and reports directly to the chief legal officer or general counsel. Technology strategy sits inside legal operations at 91 percent of those departments. A general counsel who tries to interview for “AI” without legal operations in the room is interviewing for a tool the department does not run.

The credential mix makes the two labor markets visible. CLOC’s 2025 cut of legal-operations leaders’ highest degree puts a bachelor’s at 43 percent, an MBA at 19 percent, and a project-management certification at 9 percent. That is the person who will vet the vendor, run the rollout and measure whether the tool works. The lawyer across the table still has to satisfy New York City Bar Formal Opinion 2024-5, which expressly covers corporate legal departments. Scoring both people on the same prompt-demo is how a company hires a chatbot user and still has nobody who can sign a filing.

The head of HR is in the room for a New York-specific reason. Local Law 144 of 2021, enforced from 5 July 2023, bars employers and employment agencies from using an automated employment decision tool in the city unless a bias audit was completed within one year, a summary is public, and notice is given. The Department of Consumer and Worker Protection’s FAQ states that the law applies if the job is in an NYC office, is fully remote but associated with an NYC office, or the employment agency is in New York City; that “employment decision” includes screening; that notice runs 10 business days before use; and that the employer, not the vendor, owns the audit. In the same New York interview cohort, 188 heads of talent and HR directors sitting with those legal chiefs, over a 24-month window, told Sartori the company’s own screening stack had not been scoped against Local Law 144 before the lawyer interviews began. The head of HR is therefore both a co-interviewer of the lawyer and a potential client of the same employment counsel.

Public bodies in this geo are hiring AI into policy and IT, not as a substitute for the corporate general counsel’s docket. The Department of Financial Services posted, on 12 August 2026, an Assistant Director for Policy, Responsible Artificial Intelligence Safety and Education, at $127,507 to $160,911 plus a $4,000 downstate adjustment, with a J.D. as a substitution for years of specialized experience rather than a gate. The Office of the Attorney General advertised a Director of Enterprise Artificial Intelligence, Data, and Analytics in its Information Technology Bureau at $142,032 to $149,287, reporting to the CIO, with no license required. Those bands are not the going rate for a New York in-house lawyer. They are the State splitting AI governance into DIGIT and IT while most companies are being asked, by their own chief executives, to keep it inside the legal function.

A Sedgwick survey of Fortune 500 executives, cited in a 2026 Corporate Compliance Insights essay arguing that the general counsel should lead AI strategy, found 70 percent have established AI risk committees and only 14 percent report being fully prepared for deployment. Committees without a scorecard are how an uncovered seat stays uncovered. The same empty-checklist conversation is already live for legal recruiters london, where Sartori’s interview cohort is 750 structured interviews with London lawyers.

On 22 June 2023, in Mata v. Avianca, Inc., 678 F. Supp. 3d 443, Judge P. Kevin Castel held there is nothing inherently improper about using a reliable AI tool, and that Rule 11 still imposes a gatekeeping duty to ensure accuracy. Counsel had submitted non-existent opinions with fabricated quotations and stood by them after the court ordered production. Sanctions were $5,000 jointly and severally, plus letters to the plaintiff and to each judge falsely identified as author. Every New York general counsel who lets in-house litigators, or outside counsel, file in the Southern District is already interviewing for verification literacy. Chatbot fluency is not the test. Independent cite-checking is.

The ethics opinions followed. Formal Opinion 512 in July 2024. Formal Opinion 2024-5 covering legal departments. Then the court-practice gates: there is no district-wide Southern District AI rule, and chambers of Judge Vernon S. Broderick lists a Certificate Regarding the Use of Generative Artificial Intelligence dated 29 October 2025. An in-house litigator who files in those courts needs to know which judge’s certificate is live. That is a mid-level litigation-counsel screen, not a new hire class.

Privilege made the enterprise-tool distinction non-optional. After Heppner in February 2026, a general counsel who tolerates unsupervised consumer-chatbot use on investigations or white-collar files is building a record the government can take. A 2026 Squire Patton Boggs note for general counsel on US deployment is the operational version: 2026 is about embedded and agentic AI in SaaS, HR, cyber and productivity suites, not only employee prompting. The interview is for judgment inside that estate, not for a NIST certificate. NIST’s AI Risk Management Framework 1.0 (26 January 2023) and the Generative AI Profile (26 July 2024) are voluntary language many legal departments already use to Govern, Map, Measure and Manage. ACC’s in-house framework tells departments to treat AI as enterprise risk on existing cyber, privacy and third-party architecture, with risk tiering and named business owners.

Employment counsel has been inside Local Law 144 since 5 July 2023. Penalties run $500 to $1,500 per day. Enforcement is still thin — two complaints in the Comptroller’s July 2023 to June 2025 window — which reduces an immediate surge relative to the statute’s face and does not make the duty go away. A head of talent at a listed media company told Sartori they paused the lawyer interviews for ten business days after realizing the company’s own resume screen had not been scoped as an automated employment decision tool. That pause is cheaper than a DCWP file.

At DFS-covered banks and insurers, the 16 October 2024 industry letter folds AI into existing 23 NYCRR Part 500 risk assessments, third-party, training and senior-governing-body understanding. It does not add a new AI statute. The last tranche of the Part 500 second-amendment controls, including multi-factor authentication and asset-inventory duties, has been live since 1 November 2025. The 21 May 2026 advisory, addressed to CISOs, flags frontier-AI models as a heightened cyber-threat environment and urges human oversight of AI-generated code before production. The in-house seat that answers those letters is privacy, cyber or regulatory counsel reporting to the general counsel. The CISO is the counterpart, not the substitute.

The RAISE Act is the product-and-regulatory clock. The Governor signed S6953B/A6453B on 19 December 2025; chapter amendments followed on 27 March 2026; obligations bite on 1 January 2027. A May 2026 client update on the statute describes the gate: frontier developers who develop, deploy or operate in New York models trained above 10²⁶ FLOPs; large frontier developers with more than $500 million prior-year gross revenue must publish a frontier AI framework, file DFS disclosure statements, pay assessments, and report critical safety incidents within 72 hours (24 hours if imminent death or serious injury). The Attorney General may seek civil penalties up to $1 million for a first violation and $3 million for a subsequent one. There is no private right of action. Companies that only use, license or host others’ models are not frontier developers. On the buy side, general counsel of New York financial institutions will be asked for vendor diligence that matches the published frameworks. Hire through 2026. The statute does not wait.

Public-company disclosure is already moving without a new SEC rule. The Investor Advisory Committee’s 4 December 2025 recommendation would fold AI disclosure into existing Regulation S-K items 101, 103, 106 and 303, and states that around 40 percent of companies already charge a board committee with AI oversight. Sixty-two percent of chief legal officers in the 2026 ACC survey also hold the corporate secretary function. The desk that drafts those sentences is the general counsel’s, not a new title. The Federal Trade Commission’s 25 September 2024 Operation AI Comply sweep included a proposed order against a company that claimed to sell “AI Lawyer” services ($193,000 and a ban on unsubstantiated professional-substitution claims). On 22 December 2025 the Commission set aside the Rytr consent order; that set-aside narrows one theory. It does not repeal false-claims enforcement. Advertising and consumer-protection counsel at companies making AI claims still need substantiation literacy.

New York groups that place AI on the European Union market already live under Article 4 of Regulation (EU) 2024/1689, the AI-literacy duty on providers and deployers, from 2 February 2025. That is the closest statutory analogue to this interview checklist, and it is named here for contrast. The European statutory seat is a different page. Do not run the Act from a New York scorecard.