What Published FTC Guidance, State AG Enforcement Actions, and Bar Ethics Opinions Through October 2026 Actually Say About AI-Generated Legal Intake Tools That Collect Consumer Data Before Any Attorney-Client Relationship Is Formed

Executive Summary

AI-powered legal intake tools occupy a regulatory gap that is narrowing rapidly from two directions simultaneously. Consumer protection law — anchored by Federal Trade Commission guidance, state attorney general enforcement, and comprehensive privacy statutes — treats the moment of data collection as the triggering event for regulatory obligations, regardless of whether an attorney ever reviews the intake or a client relationship ever forms. Bar ethics frameworks, by contrast, have historically oriented around the moment an attorney-client relationship begins or when a prospective client communicates with a lawyer. These two frameworks do not simply overlap; they can directly conflict, creating compounding liability exposure for law firms and legal technology vendors that deploy automated intake tools without understanding both regulatory domains.

I. The FTC Framework: Data Collection Is the Trigger

The FTC's authority over AI-driven data collection practices flows from Section 5 of the FTC Act, prohibiting unfair or deceptive acts or practices in or affecting commerce. The FTC adopted its Unfair Methods of Competition Policy Statement on November 10, 2022, which made explicit that Section 5's reach does not require a showing of market power or antitrust-style harm — it covers deceptive practices in data handling regardless of competitive context.

For legal intake tools specifically, three bodies of FTC guidance are directly relevant.

First, the FTC's longstanding enforcement position — articulated in guidance on the use of artificial intelligence and algorithms — is that automated systems collecting sensitive information must accurately disclose what data is collected, how it is used, and who receives it. The FTC has consistently treated failure to disclose AI involvement in consumer-facing systems as a potential deception violation. The DoNotPay matter is the clearest legal-sector precedent: the FTC announced its complaint and proposed consent order against DoNotPay on September 25, 2024, with the consent order finalized in February 2025. The order required $193,000 in monetary relief and prohibited advertising AI lawyer functionality without substantiating evidence. The FTC's theory did not depend on whether DoNotPay's users had formed any attorney-client relationship — the unfair and deceptive practice was the marketing claim and the data handling, full stop.

Second, the FTC's negative option marketing rule — the "click-to-cancel" amendment — was vacated by the Eighth Circuit on July 8, 2025, before its main provisions took effect. That ruling reduces one specific pressure point on intake subscription models, but it does not affect the FTC's core Section 5 authority over deceptive data-collection disclosures.

Third, the FTC Safeguards Rule imposes an independent notification obligation: entities subject to the rule must notify the FTC within 30 days of discovering a notification event affecting 500 or more consumers. Law firms handling financial data — which some intake flows do — may be within scope depending on their services.

The aggregate picture from FTC guidance is that an AI intake tool collecting names, legal matter descriptions, financial circumstances, and contact information from prospective legal clients is collecting sensitive consumer data subject to Section 5 scrutiny at the moment of collection, without any dependence on what happens afterward in the legal process.

II. State AG Enforcement: California, Texas, and New York Through October 2026

California. The California Privacy Protection Agency finalized its automated decision-making technology regulations, with the CPPA Board approving the rules on July 24, 2025, effective January 1, 2026, and ADMT compliance obligations applying from January 1, 2027. AI-driven intake systems that make, or substantially assist in making, decisions about whether to route a prospective client to an attorney, what matter category to assign, or whether to follow up likely constitute ADMT within the regulation's scope. The California AG has pursued enforcement under the CCPA: the DoorDash settlement with the California AG, finalized February 2024, resulted in a $375,000 payment. The CPPA separately settled with Honda in March 2025 for $632,500. Neither action involved legal services, but both establish that automated consumer-data collection for commercial purposes — including lead qualification — draws enforcement attention. The California AG has warned that AI systems collecting sensitive data without adequate disclosure create deceptive practice exposure under California consumer protection law.

Texas. The Texas Data Privacy and Security Act became effective July 1, 2024, with a permanent 30-day cure period. The TDPSA applies to legal services businesses that meet threshold criteria, treating sensitive data — which includes data about an individual's health, financial status, and legal proceedings — with heightened obligations. Texas Business and Commerce Code § 521.053 imposes a 60-day breach notification deadline. The Texas AG has not yet filed a publicized enforcement action specifically targeting a legal intake AI tool through October 2026, but the TDPSA's structure means that legal intake data almost certainly qualifies as sensitive personal data under Texas law, triggering consent and processing obligations regardless of whether an attorney-client relationship ever forms.

New York. New York amended GBL § 899-aa in December 2024 to add NYDFS to the list of regulators requiring breach notification for covered data. New York City's Local Law 144, administered through DCWP, defines automated employment-decision tools with notable precision: the DCWP final rules require that AI substantially assist or replace discretionary decision-making rather than merely providing supplemental information. Law firms operating intake tools in New York that use AI to screen, score, or route prospective clients should evaluate whether this definition reaches their systems. No publicized NYC LL 144 enforcement action against a legal intake tool has been filed through October 2026, but the definitional framework is in place.

III. Bar Ethics Opinions: Where the Frameworks Diverge

Published bar ethics opinions have begun addressing AI intake tools, but they approach the question from an entirely different regulatory starting point: when does a duty arise to a prospective client, and when does the collection of legal matter information through an automated system create attorney-client or prospective-client obligations?

ABA Formal Opinion 512, issued July 29, 2024, is the first ABA formal opinion on lawyers' use of generative AI. It addresses competence, confidentiality, communication, candor, supervision, and reasonable-fee duties under the Model Rules of Professional Conduct. Critically for intake tools, Opinion 512 calls for independent verification of AI output and addresses the supervision duties that apply when AI systems perform work that lawyers would otherwise perform — which intake screening plainly is. The opinion does not establish when an attorney-client relationship begins, but it makes clear that supervising lawyers bear responsibility for AI-generated communications made in the firm's name.

Florida Ethics Opinion 24-1, approved January 19, 2024, addressed competence, confidentiality, supervision, and billing implications of generative AI. It requires that informed consent be obtained before confidential client information is processed through AI tools — a requirement that creates a direct problem for intake tools, because the information is typically collected before any consent mechanism tied to legal representation has been executed.

Texas Professional Ethics Committee Opinion 705, issued February 2025, addressed generative AI use by lawyers and explicitly touched on the supervision requirements when AI tools interact with prospective clients, requiring that lawyers maintain sufficient oversight to ensure communications comply with the Rules of Professional Conduct. Under Texas disciplinary rules, information a prospective client shares during a consultation — including an AI-mediated intake — is protected even if no representation follows.

New York City Bar Formal Opinion 2024-5, issued August 7, 2024, addressed generative AI generally. The NYC Bar's framing reinforces that confidentiality obligations attach to information communicated by a prospective client, not to the existence of a formal engagement.

North Carolina 2024 Formal Ethics Opinion 1, issued November 1, 2024, addressed AI tools in law practice including intake contexts, affirming that prospective-client information carries confidentiality protection from the moment of communication.

IV. The Collision: Where the Two Frameworks Conflict

The fundamental conflict is this: consumer protection law treats the data collection event as the regulatory trigger and focuses on the entity collecting data — including legal tech vendors operating intake platforms on behalf of firms. Bar ethics frameworks focus on the attorney's duties and ask when the relationship or prospective-client communication began. Neither framework is designed to address the other.

This creates at least three collision points.

Disclosure requirements conflict with confidentiality norms. Consumer protection law — CCPA, TDPSA, and FTC Section 5 guidance — requires transparent disclosure of what AI does with collected data, including whether it is processed by third-party vendors. Bar ethics rules protect the confidentiality of prospective-client communications and constrain what can be disclosed to third-party vendors without informed consent. A firm using a third-party AI intake platform must navigate disclosure obligations that could themselves violate confidentiality if not structured carefully.

Vendor liability exists in a gap the bar rules do not cover. The FTC's theory in DoNotPay did not require an attorney-client relationship — it reached the vendor directly. Legal tech vendors operating intake tools are subject to FTC Section 5 and state AG enforcement regardless of their contractual relationship with law firms. Bar ethics rules do not reach vendors directly, but ABA Opinion 512's supervision framework means firms that deploy these tools carry residual responsibility for how vendors handle intake data.

Breach notification timelines are fixed; ethics duties are not. The California 30-day, New York 30-day, and Texas 60-day breach notification deadlines apply to intake data the moment it is collected. Bar ethics rules have no equivalent fixed timeline for notifying prospective clients of a breach of their confidential communications — the obligation is implied by competence and communication duties but is not specified. A firm that suffers a breach of intake data must comply with consumer protection notification timelines that may require public disclosure before the firm has even assessed whether a prospective-client relationship exists.

V. What Law Firms and Legal Tech Vendors Must Do Now

Several concrete steps follow from the current state of the frameworks.

Map the data at collection, not at engagement. Every field in an AI intake form — matter type, injury description, financial information, opposing party — is personal data subject to consumer protection law from the moment it is submitted. Firms and vendors should conduct data-mapping exercises that treat intake data as regulated data from the first collection event.

Align vendor contracts with supervision requirements. ABA Opinion 512's supervision framework and state equivalents require that firms maintain meaningful oversight of AI tools operating in their name. Vendor contracts should specify data handling, model behavior, retention limits, and breach notification obligations consistent with the firm's ethics duties and applicable state notification deadlines.

Implement prospective-client confidentiality protections before engagement. Texas Opinion 705, North Carolina 2024 FEO 1, and the NYC Bar's Opinion 2024-5 all support the conclusion that confidentiality attaches to intake communications before formal representation. Intake tool terms of service and privacy notices should be structured to honor this protection rather than treating prospective-client data as ordinary commercial consumer data.

Monitor CPPA ADMT compliance deadlines. California's ADMT compliance obligations take effect January 1, 2027. Firms and vendors with California users have one compliance cycle to build opt-out mechanisms and conduct required impact assessments for intake systems that qualify as ADMT.

The window to treat AI legal intake tools as a purely internal technology decision closed when FTC enforcement reached DoNotPay and state comprehensive privacy laws became effective. The firms and vendors that treat the intake data event as a dual-regulated moment — governed simultaneously by consumer protection law and professional responsibility rules — are the ones positioned to avoid enforcement on both tracks.