What it does

S.7263 / A.6545 would add a new section — § 390-f — to Article 22-A of the New York General Business Law (the article that governs consumer-protection causes of action). The bill targets a single thing: chatbots that hold themselves out as, or behave like, members of state-licensed professions — lawyers, doctors, dentists, psychologists, social workers, architects, engineers, and the rest of the professions whose licensure is governed by Article 130 of the New York Education Law and Article 15 of the Judiciary Law².

The bill creates four operative requirements:

1. A prohibition. An “operator” (anyone who owns, operates, or deploys a chatbot) cannot let the chatbot give “substantive responses, information, or advice, or take any action” that — if a human did it — would be the unauthorized practice of a covered profession or unauthorized use of a professional title¹.
2. An imputed-knowledge rule. The operator “shall be deemed to know” that a chatbot cannot lawfully practice a covered profession. There is no “we didn't know” defense².
3. A no-disclaimer-by-disclosure clause. The operator “may not waive or disclaim this liability merely by notifying consumers that they are interacting with a non-human chatbot system.” The standard “I'm an AI assistant, not a doctor” banner does not immunize the operator³.
4. A notice requirement. Operators must still provide “clear, conspicuous and explicit notice” that users are interacting with an AI, in the same language and font size as other prominent text on the site².

Enforcement runs along two tracks. The Senate version creates a private right of action: anyone harmed can sue for actual damages, and willful violations trigger attorneys' fees and litigation costs³. The Assembly version (current print: A.6545b) adds an Attorney General enforcement track: civil penalties up to $15,000 per violation, plus a public AG-run complaint portal². Both versions would take effect 90 days after enactment.

The operative statute

The bill's core is one new statutory section. Below is what § 390-f would say once enacted, paraphrased to the level of detail public reporting allows²:

§ 390-f. Liability for chatbots impersonating licensed professionals.

Definitions. “Chatbot” means an artificial intelligence system, generative artificial intelligence, software program, or technological application that simulates human-like conversation in response to user inputs. “Operator” (Senate uses “proprietor”) means any person or entity that owns, operates, or deploys such a chatbot, excluding third-party developers who merely license the underlying model to the operator. “Covered profession” means a profession licensed under Title VIII of the Education Law or Article 15 of the Judiciary Law.

Prohibition. No operator shall permit a chatbot to provide substantive responses, information, advice, or take any action which, if performed by a natural person, would constitute the unauthorized practice of a covered profession or unauthorized use of a professional title in violation of Education Law §§ 6512 or 6513 or Judiciary Law §§ 478–485.

Imputed knowledge. An operator shall be deemed to know that a chatbot cannot lawfully practice a covered profession or appear as an attorney-at-law.

No waiver by disclosure. An operator may not waive or disclaim liability under this section merely by notifying users that they are interacting with a non-human chatbot system. The operator must nonetheless provide clear, conspicuous, and explicit notice of the chatbot's artificial nature in the same language and font size as other prominent text on the platform.

Remedies. A person injured by a violation may bring a civil action to recover actual damages; on a finding of willfulness, the court shall award attorneys' fees and reasonable costs of litigation. The Attorney General may seek injunctive relief and civil penalties of up to fifteen thousand dollars per violation, and shall maintain a public-facing complaint portal regarding violations of this section.

Effective date. Ninety days after this act becomes law.

Two definitional points carry most of the bill's force:

• Who is an operator. Importantly, the definition excludes third-party developers who license their underlying model to a deployer. The model providers themselves — OpenAI, Anthropic, Google, Meta, and others — are not direct targets of the bill. The integrator (the company embedding a chatbot in their platform) is³.
• Who is a covered profession. The bill specifically enumerates 13 professions licensed under Article 130 of the Education Law — medicine, psychology, mental-health services, veterinary medicine, dentistry, physical therapy, pharmacy, nursing, podiatry, optometry, engineering, architecture, and social work — plus the unauthorized practice of law under Article 15 of the Judiciary Law. Fourteen covered professions in total¹⁵.

Why it matters

The bill is the first U.S. attempt I am aware of to assign civil liability for unauthorized practice of a profession to AI deployers in a clean, statutory, plaintiff-friendly form. Two existing frameworks have skirted the issue:

• The federal route (FTC). In September 2025 the Federal Trade Commission ordered seven chatbot companies to produce safety reports, but the FTC's authority runs through deception and unfair-practice doctrine, not through state professional-licensing law⁴.
• The state-AG-guidance route (California). The California Attorney General published guidance in 2025 stating that “California law does not allow delegation of the practice of medicine to AI” and that licensed-physician supervision is required, but that guidance is interpretive of existing medical-practice law, not a new statutory liability⁵.

S.7263 / A.6545 takes a different swing. Rather than relying on the FTC's deception authority or on existing professional-practice statutes (which were drafted decades before chatbots and have varying applicability), it creates an express civil cause of action with an explicit imputed-knowledge rule and an explicit no-waiver clause. That is structurally novel.

“You can't get away with it by simply slapping a disclaimer on your chatbot that it's AI.”Elena Gurevich, AI attorney, on S.7263 (PPC.land, March 2026)³

Senate vs. Assembly: the bills are diverging

Same-as bills usually march in lockstep. These two have not. Several substantive differences have emerged across amendment cycles — and the Senate version is meaningfully further along procedurally, having been unanimously voted out of the Senate Internet & Technology Committee on February 25, 2026 as part of an 11-bill chatbot package¹¹.

The two versions are likely to converge before either could be enacted — same-as bills must match for delivery to the Governor. The Assembly's bipartisan co-sponsor list and the addition of AG enforcement suggest the Assembly version is the politically negotiated one; if the bill moves, the Assembly's shape is the more likely final shape.

Hearings & testimonyAdded

The bill has not had its own dedicated hearing — but it has been heard in the context of two consequential Senate Internet & Technology Committee proceedings in 2026.

January 15, 2026 — Public hearing on AI in high-risk contexts

Sen. Gonzalez, as committee chair, convened a full-day public hearing in Albany titled “Public Hearing: To discuss risks, solutions, and best practices with respect to the use of artificial intelligence in consequential or high-risk contexts.” She was joined by Senators Hinchey, Andrew Gounardes, Jabari Brisport, John C. Liu, Nathalia Fernández, Julia Salazar, Kevin Parker, and Gustavo Rivera, and by Assembly Members Michaelle Solages (Assembly sponsor of the New York AI Act, S.1169A) and Steve Otis (chair, Assembly Science & Technology Committee — the committee that now holds A.6545)¹².

The witness list spanned labor, civil society, academia, and industry:

• Labor: Henry Garrido (Executive Director, DC 37 — NYC's largest municipal-employee union) and Mia McDonald (Senior Campaign Lead, Communications Workers of America)¹³.
• Civil society / academia: Kara Williams (Electronic Privacy Information Center, EPIC); academics from NYU, Haverford College, and others¹³.
• Industry & trade: Chamber of Progress (the tech-industry trade group; submitted written testimony arguing that “AI has tremendous potential for improving education, enabling creative expression, and creating new business opportunities, and stress[ing] that it is critically important that public policy promotes the broad and equitable distribution of these innovations”); the Business Council of New York State (the state's largest employer association, representing 3,400 private-sector employers); the Innovative Lending Platform Association¹².

Source: NY Senate Internet & Technology Committee public hearing, January 15, 2026; witness list and written-testimony archive at nysenate.gov.

The hearing's framing was the broader New York AI Act (S.1169A / Assembly companion), but witnesses were directly asked about the chatbot-impersonation question that S.7263 / A.6545 addresses. There is no committee transcript published yet; the witness list and most written testimony are publicly archived as PDFs on the Senate site¹².

February 25, 2026 — Senate Internet & Technology Committee passes the chatbot package

“On Wednesday, February 25, 2026 the New York State Senate Committee on Internet & Technology, Chaired by State Senator Kristen Gonzalez, passed a package of legislation to regulate AI chatbots… including legislation to protect minors from unsafe features in chatbots [S.9051] and to prevent AI from impersonating certain licensed professionals, including mental healthcare workers [S.7263].”

— NY Senate press release, February 26, 2026¹¹

The committee passed S.7263 unanimously. The press release explicitly highlights mental-health-worker impersonation — an editorial choice that signals the politically resonant case for the bill: AI chatbots that hold themselves out as therapists. (The American Psychological Association warned the FTC in early 2025 that AI chatbots were “masquerading” as licensed therapists; Holland & Knight's legal-industry analysis later identified that warning as the bill's likely motivating event)¹⁴.

Voices on the record

Two on-the-record statements from the broader 2026 AI-package debate are particularly relevant to S.7263 / A.6545's politics:

“AI is… not a replacement for human judgment or jobs.”

— Mario Cilento, President, New York State AFL-CIO, on the chatbot-regulation package¹⁵

AI “threatens to deepen economic inequality and destabilize our economy by eliminating entire job sectors overnight.”

— Sen. Michelle Hinchey (D-Kingston, Working Families Party), co-sponsor of S.7263¹⁵

Both quotes locate the bill in a labor-protective framework. That framing matters because it indicates the political coalition that will carry the bill: organized labor (DC 37, CWA, NYS AFL-CIO), professional licensure interests, and the progressive bloc within the Senate Democratic caucus. The Assembly's bipartisan ~37 co-sponsors suggests broader buy-in there — including consumer-protection-minded Republicans willing to back what they read as a fraud-prevention measure.

Legislative timeline

Neither chamber has yet held a substantive public hearing on the bill. The Internet & Technology Committee's recent activity has been dominated by Sen. Gonzalez's separate S.9051 (the AI Chatbot Ban for Minors, developed with AG Letitia James's office and Common Sense Media), which passed committee 7-0 on Feb. 25, 2026 and advanced to Senate Finance⁷.

The wider AI-regulation map

S.7263 / A.6545 sits in the middle of the most crowded U.S. AI-regulation moment to date. Roughly the past 18 months:

The pattern across all of these: every other recent bill or rule targets either (a) child safety / mental health, or (b) generic disclosure ("you are talking to an AI"). S.7263 / A.6545 is the only one in the set that hits the professional-services use case head-on, with a specific statutory cause of action and an explicit no-waiver-by-disclosure rule.

Administrative and case law

The bill operates by reference to existing professional-licensure law. Three layers of statute and regulation are doing the substantive work behind it.

The Education Law backbone

New York Education Law Article 130 governs the licensing of professions other than law. Education Law § 6512 makes the unauthorized practice of a licensed profession a Class E felony; § 6513 covers unauthorized use of a professional title (Class A misdemeanor). Together these statutes are the criminal floor that S.7263 / A.6545 imports into civil law⁹. The relevant agency is the New York State Education Department's Office of the Professions, which licenses and disciplines practitioners under regulations in Title 8 of the New York Codes, Rules and Regulations (NYCRR).

The Judiciary Law backbone

For the unauthorized practice of law, New York Judiciary Law §§ 478–485 forbid non-attorneys from practicing law or holding themselves out as attorneys. Enforcement is concurrent: the Attorney General can prosecute, district attorneys can prosecute, and the Appellate Division can issue contempt orders. The bill imports this framework as well⁹.

What case law currently applies to AI “practice”

There is, as of this writing, no New York appellate decision applying the Education Law or Judiciary Law unauthorized-practice statutes to a chatbot or other AI system. The closest cases are pre-AI: El Gemayel v. Seaman, 72 N.Y.2d 701 (1988) (foreign attorney advising in NY); Spivak v. Sachs, 16 N.Y.2d 163 (1965) (out-of-state lawyer). These define unauthorized-practice doctrine but predate any AI-specific application¹⁰.

The bill is therefore a statutory clarification: rather than waiting for a court to extend Article 130 / Judiciary Law unauthorized-practice doctrine to AI — which would be plausible but uncertain — it nails the question down and assigns the cause of action explicitly. That is the same structural move the Council made with NYC LL 64 of 2023 (statutorily ratifying what would otherwise be administrative): convert “courts might do this” into “the statute does this.”

Federal preemption questions

One open legal question: whether federal AI policy — in particular any future federal preemption clause inserted into an executive order, FTC rule-making, or congressional action — would override the New York statute. As of May 2026 there is no federal statute on point that would clearly preempt; the FTC's Section 6(b) authority operates alongside state consumer-protection law, not in place of it⁴.

The critique

The bill has drawn substantive policy criticism that deserves to sit alongside the operative analysis. The most-circulated version came from George Mason economist Alex Tabarrok at Marginal Revolution, who in March 2026 published a post titled “Claude on NY's Senate Bill S7263” (a critique generated using the Claude AI assistant)⁸. Three lines of objection:

The legal-ratchet objection

The bill says a chatbot can't do anything that would be a crime if a human did it. But the criminal threshold for unauthorized practice is high — a friend with a JD can talk you through your lease without committing a crime; a retired physician can tell you what your rash looks like without committing a crime. So either the bill barely bites (because the human equivalent is rarely criminal), or courts will read it expansively and create a de facto stricter rule for AI than for humans. Either reading produces unintended results⁸.

The over-deterrence objection

Faced with ambiguous liability, operators will systematically over-restrict their chatbots' outputs — refusing to engage with anything that even brushes against medicine, law, or accounting, regardless of how anodyne the actual question is. The compared baseline matters here: it isn't “AI advice vs. licensed-professional advice” for most users facing an eviction notice or a tax letter; it's “AI advice vs. no advice at all.” The people most likely to lose access are the people who cannot afford the licensed alternative⁸.

The public-choice / capture objection

Tabarrok frames the bill in classic public-choice terms: organized professional associations secure statutory protections framed as consumer safeguards; the costs are dispersed across millions of unorganized users who “will never know what information they didn't receive.” This is the structural form of regulatory capture identified in Buchanan and Tullock's 1962 work and applied by economists for sixty years to occupational licensing⁸.

Other critical voices

Tabarrok's analysis is the most-circulated, but several others have emerged in legal-industry and policy outlets:

• Holland & Knight (Nili Yolin and Shalyn Watkins, March 6, 2026) call the bill “a significant but potentially problematic regulatory experiment” and predict it “may do less to meaningfully curb unlicensed professional conduct than expand litigation exposure for chatbot deployers.” They flag a “Disclosure Paradox”: the bill requires the AI-disclosure notice but explicitly forbids it from immunizing the operator — creating “false compliance expectations”¹⁴.
• Underberg & Kessler (Jacob H. Zoghlin, March 10, 2026) emphasize an angle others have missed: municipal exposure. The bill's “proprietor” definition explicitly includes government entities. NYC, county governments, and small municipalities running chatbots for permits, building regulations, or property guidance face the same liability as commercial deployers¹⁶.
• Reason (Jack Nicastro, March 4, 2026) frames the bill as “protectionist legislation masquerading as consumer protection” and quotes Kevin Frazier of the Cato Institute calling it “unduly limit[ing] access to information in a manner that is not only unconstitutional, but also contrary to both democratic values and a free market economy” — comparing the restriction to demanding libraries remove legal and mental-health resources. Reason also quotes Taylor Barkley of the Abundance Institute calling the bill “shortsighted at best and protectionist at worst”¹⁷.

The opposing view — on the record from labor, advocates, and the bill sponsors — is that the consumer-protection problem is real: chatbots do sometimes give confidently wrong professional-grade advice (the American Psychological Association formally warned the FTC in early 2025 about AI chatbots “masquerading” as therapists)¹⁴, and the population most likely to rely on a chatbot's medical or legal “answer” without verification is the same low-resource population the over-deterrence objection invokes. Whose risk dominates is an empirical question the bill's text doesn't try to answer.

Politics & likely pathAdded

The bill is in the unusual position of being well-advanced in one chamber, slow-moving in the other, and ideologically hard to oppose head-on — which makes its eventual fate easier to handicap than most pending legislation.

Who's pushing it: the labor-and-progressives axis

The Senate version's coalition is narrow but cohesive: Sen. Gonzalez as committee chair drives it; Sens. Hinchey, Liu, and Salazar co-sponsor; the broader 2026 AI package was endorsed by NYS AFL-CIO President Mario Cilento and witnessed by DC 37's Garrido and CWA at the January 15 hearing¹⁵. This is the same labor-progressive coalition that backed Hinchey's parallel Workforce Stabilization Act (which would require labor-impact assessments before AI deployment and impose a worker-displacement surcharge)¹⁵.

The Assembly version's coalition is much broader. Asm. Zaccaro Jr.'s ~37 co-sponsors include Republicans (Yeger, Ra, Sempolinski, Norber, Gallahan, Manktelow, Pirozzolo, Chludzinski, McDonough, Novakhov, Miller, K. Brown, Gray) alongside Democrats spanning the ideological range. That breadth is unusual for a tech-regulation bill in NY and reflects the framing as fraud-prevention rather than industrial regulation².

Who opposes it

Direct, on-the-record opposition has been measured. Three identifiable currents:

1. Tech industry & venture-capital lobby. A May 29, 2025 letter from venture-capital partners to Senate Majority Leader Andrea Stewart-Cousins and Speaker Carl Heastie opposed the broader RAISE Act (the AI safety/transparency bill that passed in heavily-amended form in 2025) on innovation-harm grounds. The same lobby has been active against Gonzalez's broader New York AI Act (S.1169A); their posture on the narrower S.7263 has been less public but is implicit in Chamber of Progress's testimony at the January hearing¹⁸.
2. Free-market policy organizations. The Cato Institute (Kevin Frazier) and the Abundance Institute (Taylor Barkley) have publicly criticized the bill as protectionist and access-restricting. Reason magazine has carried the most aggressive coverage¹⁷.
3. Big employer associations. The Business Council of New York State (3,400 private-sector members) testified at the January 15 hearing on AI regulation broadly. Their substantive position on S.7263 specifically is not in the public press releases but their default posture — opposing new private rights of action with attorneys' fees — is implicit¹².

Notably absent from the opposition: any organized professional association. The medical, legal, mental-health, dental, engineering, and architecture lobbies have not publicly opposed the bill. They are its silent beneficiaries — which is exactly Tabarrok's public-choice point⁸.

Hochul's complicated AI posture

Governor Kathy Hochul's record on AI legislation is the central variable in the bill's fate. The pattern over 18 months:

The pattern: Hochul will sign AI-regulation bills, but typically after they've been amended to reduce industry exposure. She has actively used both the bully pulpit and the budget to position New York as “a leader in AI research” while also “highlighting the dangers of deceptive generative AI”¹⁹. She has not publicly weighed in on S.7263 specifically.

Likely path, in three scenarios

Each scenario assumes the bill clears the Senate floor in the 2025–26 session. The Senate version is calendar-ready; the question is whether and how it harmonizes with the Assembly.

Factors that will swing the outcome

• Federal preemption pressure. If Congress moves on the CHATBOT Act or GUARD Act in 2026 or 2027 with broad preemption language, NY's bill becomes harder to defend procedurally even if substantively unchallenged⁷.
• A high-profile incident. The 2025 spate of news stories about chatbots giving harmful medical or mental-health advice is what pushed the bill onto the calendar in the first place. Another incident — particularly one involving a minor or a death — would likely accelerate enactment in toughened form.
• AG Letitia James's posture. Her office co-developed S.9051 (the chatbots-and-minors bill); a public endorsement of S.7263 would meaningfully change the political weight on Hochul's desk. Without it, the bill remains a Gonzalez-driven measure rather than an executive-branch priority.
• Industry restructuring announcements. If a large legal-tech or telehealth platform publicly announces NY service restrictions in response to the bill, expect rapid amendment to add carve-outs. If no such announcement comes, expect the bill to pass closer to current form.

Where things stand now

As of May 5, 2026:

• Senate: S.7263 has been on third reading since May 7, 2025 and was advanced again March 4, 2026. It reached the Senate calendar Feb. 26, 2026. A floor vote could come at any time once leadership schedules it.
• Assembly: A.6545 has been recommitted three times with successive amendments; current print is 6545b (May 4, 2026). It now sits in Science and Technology after being reassigned out of Consumer Affairs and Protection in March 2026. Bipartisan co-sponsor list (~37 names spanning both parties) suggests the bill could clear the Assembly if scheduled.
• The bills are not aligned. The Assembly's enforcement provisions (AG track, $15K civil penalty, complaint portal) are absent from the Senate version. Same-as bills must match before delivery to the Governor; one chamber will need to take the other's amendments.
• No public hearings yet. Neither chamber has held an extended hearing on the bill specifically. The committee record consists of referrals and amendments, not testimony.
• No companion executive action. Neither AG Letitia James nor Governor Hochul has formally weighed in on this specific bill, though the AG's office has co-developed Sen. Gonzalez's separate AI-and-minors bill (S.9051), suggesting a sympathetic posture⁷.

What we still don't know

• The actual industry response. Will major chatbot deployers (legal-tech platforms, healthcare-app companies, mental-health chatbots, financial-advice tools) restructure their NY exposure if the bill passes? Restructuring options include: route NY users to licensed humans; geofence professional-domain conversations; insert binding arbitration; contractually shift liability onto integrators. None of these have been publicly announced.
• Whether AG James will publicly support the bill. Her office's involvement in the Gonzalez/AI-and-minors bill suggests she is favorably disposed to AI consumer-protection legislation. Her formal endorsement of S.7263 / A.6545 has not been recorded.
• Whether Governor Hochul will sign if both chambers pass it. She has signed other consumer-protection technology measures and has not publicly opposed AI-liability legislation.
• How “substantive responses, information, or advice” will be construed. The phrase is the bill's main interpretive battleground. Drug-interaction warnings on a pharmacy chatbot? Tax-deduction estimators on an accounting platform? Property-line questions on a real-estate site? All plausibly within reach; all currently unsettled.

Prior Manhattan Institute workAdded

The Manhattan Institute's AI-policy work has been published primarily on its main site and in City Journal, with broader anti-licensing-restriction work spanning two decades. The Institute's general posture on AI regulation is innovation-friendly and federalist-skeptical of state patchworks — which puts it on the critical side of S.7263 / A.6545, though without published commentary on this specific bill (yet). The most directly relevant pieces:

The case against state-by-state AI regulation

• “Blue State AI Regulations Could Threaten U.S. Competitiveness” — City Journal (July 25, 2025). The Institute's headline statement of position: “more than 1,000 AI-related state and local laws are now pending nationwide, risking a patchwork of costly and confusing regulations.” Notes that one 2024 survey of state proposals identified 57 different definitions of “artificial intelligence” or “automated decision system” — directly relevant to S.7263's broad “chatbot” / “substantive response” definitional ambiguity flagged by Holland & Knight in this brief. city-journal.org →
• “The White House's AI Strategy Is Too Little, Too Late” — City Journal (March 24, 2026). Argues for federal preemption of state AI rules — the same federalist concern that figures in this brief's "factors that swing the outcome" subsection. city-journal.org →
• “On AI, More Understanding, Less Regulation” — City Journal (August 1, 2024). The general MI line: that regulatory rush gets ahead of technical understanding. city-journal.org →
• “A Conflict of AI Visions” — City Journal. Frames the federalist-vs-state-regulation debate as a contest between two distinct theories of how AI's risks materialize. city-journal.org →

AI-policy primers

• “A Playbook for AI Policy” — Manhattan Institute report by fellow Nick Whitaker. Primer on AI development history and four principles for guiding AI policy: methods for evaluating AI capability, controlling AI systems, the role of AI agents, and global competition. The Institute's most systematic statement of position. manhattan.institute →
• “Balancing AI Innovation with National Security” — Manhattan Institute. Frames AI regulation as a national-security question and argues against rules that constrain U.S. firms while leaving foreign competitors unconstrained. manhattan.institute →
• “New Report Finds Bias in ChatGPT” — Manhattan Institute. Empirical work on political/demographic bias in chatbot outputs — relevant to the consumer-protection framing that motivates S.7263 but pointed in a different direction. manhattan.institute →
• “AI Can Help Struggling Students” — City Journal. The educational-equity case for AI access — directly relevant to the “AI advice vs. no advice at all” argument in this brief's critique section. city-journal.org →

The longer line on occupational licensing

The bill's structural form — importing professional-licensure law into a new statutory cause of action — runs into a critique the Institute has been making for two decades:

• “Occupational Licensing Reduces Interstate Mobility” — Manhattan Institute. Foundational finding: people in state-specific licensed occupations have an interstate migration rate 36% lower than other workers. The MI version of the public-choice critique that Tabarrok applies to S.7263 in this brief's critique section. manhattan.institute →
• “How Occupational Licensing Harms the Young” — Manhattan Institute. Distributional analysis: the costs of licensing fall hardest on young, low-income, and minority workers. manhattan.institute →

The new live precedent: Pennsylvania v. Character.AI

This precedent post-dates the Manhattan Institute's published commentary on S.7263 but is directly load-bearing for the bill's legal architecture. On May 1, 2026, the Pennsylvania Department of State sued Character Technologies, Inc. in Commonwealth Court — alleging that a Character.AI chatbot named “Emilie” held itself out as a licensed psychiatrist with a fabricated Pennsylvania license number, in violation of state medical-practice statutes. The case underscores that regulators are already willing to apply existing medical-practice law to AI — without any AI-specific statute, which is exactly the question S.7263 is structured to settle. Sources covering the suit: NPR; Washington Times; CBS News; The Hill. Lexology has paired analyses: “No AI Law, No Problem” and “Pennsylvania Targets AI Chatbot for the Unauthorized Practice of Medicine”.

The Pennsylvania action complicates S.7263's positioning. The case demonstrates that existing professional-practice statutes can reach chatbot impersonation without new legislation; the question S.7263 / A.6545 actually decides is whether to add a private right of action on top of (or in place of) the existing AG-enforcement track that Pennsylvania is using.

The Manhattan Institute has not yet published commentary on either S.7263 or the Pennsylvania action. Given the Institute's positioning on AI regulation (federalist-skeptical, innovation-favoring, licensing-skeptical) and on occupational licensing (broadly critical), one would expect any forthcoming commentary to land in the Tabarrok / Cato Institute / Abundance Institute camp documented in this brief's critique section — though the Pennsylvania case has shifted the question from “should AI face liability?” to “through which procedural channel?”.