New York

A large developer of a frontier AI model must implement a written safety and security protocol before deploying the model, retain an unredacted copy of it for as long as the model is deployed plus five years, conspicuously publish a redacted copy and transmit that copy to the Attorney General and the Division of Homeland Security and Emergency Services, and grant either of them access to the protocol on request.

It must record the specific tests and test results behind any assessment the section or its own protocol requires, in enough detail for a third party to replicate the testing procedure, and retain those for the same period.

It must not deploy a frontier model that would create an unreasonable risk of critical harm, must review its protocol annually and republish it if that review makes a material modification, must disclose each safety incident to the Attorney General and the Division within 72 hours, and must not knowingly make false or materially misleading statements in the documents the section produces.

The Attorney General may bring a civil action for a civil penalty of up to $10,000,000 for a first violation and $30,000,000 for any subsequent violation, or for injunctive or declaratory relief; the act creates no private right of action. Signed December 19, 2025 as chapter 699 of the Laws of 2025, and in force since March 19, 2026, the ninetieth day after enactment.

Note that the enacted print, S6953-B, carries no third-party audit: the annual independent audit of compliance, its report contents and its publication duty appeared in the original print and in S6953-A and were dropped before passage, and the word audit does not appear in the chaptered text.

A depicted individual has a civil cause of action against a person who discloses, disseminates or publishes sexually explicit material showing them that was created or altered through digitization (defined to include software, machine learning, artificial intelligence, or other computer-generated or technological means) without their consent, where the defendant knew or should have known that consent was lacking.

A disclaimer stating the depiction is unauthorized or fabricated is not a defense, and the finder of fact may award injunctive relief, punitive damages, compensatory damages, and attorney's fees.

A person who uses a deceased performer's digital replica, a newly created, computer-generated, highly realistic electronic representation of their voice or visual likeness, in an audiovisual work, sound recording, or live musical performance, knowing the use was unauthorized by the applicable right holder, is liable for the greater of $2,000 or the injured party's compensatory damages, plus disgorgement of the defendant's profits and, in the court's discretion, punitive damages.

The section exempts parody, satire, commentary, documentaries and similar works, and news, public affairs and political-campaign uses. The digital replica provisions were first enacted effective 23 December 2022 and were further revised, per the consolidated law page's own revision history, effective 19 December 2025.

It is a class A misdemeanor to intentionally disseminate or publish an intimate image of another person without consent and with intent to cause emotional, financial or physical harm, including an image created or altered by digitization; the statute defines digitization to mean altering an image in a realistic manner using images of a person other than the person depicted, or computer-generated images.

The offense applies regardless of the actor's role in creating the original image, so a person who disseminates an AI-generated or AI-altered intimate depiction is exposed on the same terms as one who disseminates an unaltered photograph.

An operator of an AI companion, a system using artificial intelligence, generative AI or emotional-recognition algorithms to simulate a sustained human-like relationship with a user, must provide a clear and conspicuous notification at the start of an interaction and at least every three hours that the user is not communicating with a human, and must implement a protocol reasonably designed to detect and respond to a user's expressions of suicidal ideation or self-harm by referring them to crisis services such as the 9-8-8 hotline.

The Attorney General may enjoin a violation and seek civil penalties of up to $15,000 per day per violation of the notification or self-harm-detection duties. The article defines the operator bound by these duties as any person, partnership, association, firm or business entity that operates for or provides an AI companion to a user.

A person, firm, association, corporation, campaign, committee or organization that distributes or publishes a political communication produced by or including materially deceptive media, defined to include content created by software, machine learning, artificial intelligence or other computer-generated or technological means that is not distinguishable from reality to a reasonable person and depicts a scenario that did not occur, must disclose that use when it has actual knowledge the media is materially deceptive.

The disclosure must read "This (image, video, or audio) has been manipulated" in a specified legible format, subject to exemptions for satire, parody and bona fide news reporting. A depicted candidate may seek expedited injunctive relief, court costs and attorney's fees.

A person who, for the purpose of a sale or advertisement of property or a service, produces or creates an advertisement containing a synthetic performer, an audio, visual or audiovisual performance of a human performer not recognizable as any identifiable natural performer, must conspicuously disclose that a synthetic performer is in the advertisement where they have actual knowledge of that fact.

A violation carries a civil penalty of $1,000 for a first violation and $5,000 for any subsequent violation.

The law exempts advertisements for expressive works such as motion pictures, television programs, and video games, consistent with the use of the synthetic performer in the underlying work, and does not limit rights of privacy or publicity under Civil Rights Law sections 50 and 50-f. Signed December 11, 2025 as chapter 617 of the Laws of 2025, and in force since June 9, 2026, 180 days after enactment.

Any person or business that owns or licenses computerized data including private information must disclose a breach of the security of the system to each affected New York resident in the most expedient time possible and without unreasonable delay, and no later than 30 days after the breach is discovered.

Notice to the Attorney General, the Department of State, and the Division of State Police is required for every disclosure, and notice to nationwide consumer reporting agencies is required when more than 5,000 New York residents are notified at once.

'Private information' is personal information combined with an unencrypted data element such as a Social Security number, driver's license number, financial account number, biometric information, or medical or health insurance information; biometric information here carries no exclusion for data derived from a photograph, video, or audio recording, unlike a comprehensive-regime state's biometric definition, though this plain-text reading has not been tested in New York case law or Attorney General guidance.

This section's local-law preemption clause is scoped to breach notification and does not displace New York City's separate biometric-privacy ordinance, which regulates capture consent and retention, a different subject.

The Attorney General may bring an action to enjoin and restrain a violation of SHIELD's breach-notification duty, and a court may award actual costs for a failure to notify and, for a knowing or reckless violation, a civil penalty of the greater of $5,000 or up to $20 per instance of failed notification, capped at $250,000; a three-year limitations period runs from Attorney General discovery or notice, extendable to six years if the breach was concealed.

A violation of the separate data-security-program duty is deemed a violation of General Business Law section 349, letting the Attorney General bring an action to enjoin it and obtain civil penalties on that basis.

Section 899-bb expressly bars a private right of action for the safeguards duty, and section 899-aa carries no comparable express bar in its own text, but its enforcement subdivision is written entirely in terms of Attorney General authority, with no private-suit provision found anywhere in that section. Neither half of SHIELD arms a private plaintiff.

Section 156.00(8) defines without authorization to mean using or accessing a computer, computer service, or computer network without the permission of the owner or lessor, or someone licensed or privileged by the owner or lessor, where such person knew that his or her use or access was without permission or after actual notice to such person that such use or access was without permission, an express, textual revocation rule under which a communicated notice, including a cease-and-desist letter, that access is unwanted converts continued access into without authorization access, without needing a technical block at all.

The same subsection also provides that the knowing use of a set of instructions, code, or computer program that bypasses, defrauds, or otherwise circumvents a security measure installed or used with the user's authorization shall be presumptive evidence that such person used or accessed such computer without authorization, giving circumvention its own statutory evidentiary weight.

Section 156.05 (unauthorized use of a computer, a class A misdemeanor) and section 156.10 (computer trespass, a class E felony, requiring either intent to commit a felony or knowing access to computer material) build on the same authorization definition.

Register.com, Inc. v. Verio, Inc. (356 F.3d 393, 2d Cir. 2004), a New York-originated dispute, affirmed a preliminary injunction on a trespass-to-chattels theory where an automated WHOIS-query tool kept accessing a database after repeated exposure to posted terms of use prohibiting that use, even though the defendant never signed an agreement, holding repeated notice sufficient.

National Basketball Association v. Motorola, Inc. (105 F.3d 841, 2d Cir. 1997), applying New York's hot-news misappropriation tort, held the tort survives Copyright Act preemption only narrowly (the plaintiff generates time-sensitive information at a cost, the defendant free-rides on it, the parties are direct competitors, and free-riding would substantially threaten the product's existence or quality), and on the facts found no violation because Motorola independently gathered its own underlying facts rather than copying the NBA's compiled product; a scraper that copies a compiled output rather than independently gathering facts could still fall within the surviving tort.

No reported New York case squarely applies section 156.00 to a plain scraping fact pattern with no notice and no technical barrier.

The ordinance applies only to a commercial establishment, a place of entertainment, retail store, or food and drink establishment. Section 22-1202(a) requires clear, conspicuous signage disclosing the collection, retention, conversion, storage, or sharing of biometric identifier information, defined to include a retina or iris scan, fingerprint, voiceprint, or hand or face geometry scan, or other identifying characteristic.

Section 22-1202(b) flatly bans selling, leasing, trading, sharing in exchange for anything of value, or otherwise profiting from biometric identifier information.

Section 22-1203 creates a private right of action: an aggrieved person may sue, with a 30-day notice-and-cure period required before suing over a signage violation but no notice-and-cure required for a sale or profit violation under section 22-1202(b), and a prevailing party may recover statutory damages, attorneys' fees and costs, and injunctive relief.

It is a municipal law, narrower in scope than a statewide biometric statute because it is limited to commercial establishments' customer-facing collection and does not apply outside New York City, and its private right of action is stated in the code text. Section 22-1204 exempts government agencies entirely, and exempts financial institutions and non-analyzed photo or video capture from the disclosure duty specifically.

Any person or business that owns or licenses computerized data including the private information of a New York resident must develop, implement, and maintain reasonable safeguards to protect the security, confidentiality, and integrity of that private information.

A business satisfies this duty either by being a compliant regulated entity under Gramm-Leach-Bliley Act Title V regulations, Health Insurance Portability and Accountability Act (HIPAA) and HITECH regulations, or 23 NYCRR Part 500, or by implementing a data security program with reasonable administrative, technical, and physical safeguards scaled to the size of the business.

Failure to comply is deemed a violation of General Business Law 349, enforceable only by the Attorney General for injunctive relief; the section creates no private right of action. General Business Law 350-d sets the civil penalty for that violation at up to $5,000 per violation.

Section 500.17 of 23 NYCRR Part 500 requires a Covered Entity to notify the Superintendent of Financial Services electronically as promptly as possible but no later than 72 hours after determining that a Cybersecurity Incident has occurred at the covered entity, its affiliates, or a third-party service provider.

A Cybersecurity Incident is a Cybersecurity Event, a broader term covering any attempt to gain unauthorized access to or misuse an information system, that also either requires notice to a government or supervisory body, carries a reasonable likelihood of materially harming a material part of normal operations, or results in the deployment of ransomware within a material part of the covered entity's information systems.

Each covered entity must promptly provide the Superintendent with any information the Superintendent requests about a reported incident and must continue updating the Superintendent with material changes or new information, with no separate numbered clock on either duty.

Where a covered entity makes an extortion payment in connection with a cybersecurity event, it must notify the Superintendent of the payment within 24 hours of making it and, within 30 days of the payment, provide a written description of why the payment was necessary and what alternatives and compliance diligence it performed.

By April 15 each year, a covered entity must submit either a written certification that it materially complied with this Part for the prior calendar year or a written acknowledgment identifying the sections it did not materially comply with and a remediation timeline, signed by its highest-ranking executive and its Chief Information Security Officer.

The limited exemptions in 23 NYCRR 500.19(a), (c) and (d), covering a small covered entity and an entity without its own information systems or nonpublic information, each work by naming the sections they relieve, and none of them names section 500.17, so an entity holding one of those still owes the Superintendent-notice duty in full.

Subsections 500.19(b), (e) and (g) work the other way, exempting their named classes from the requirements of the whole Part, which carries section 500.17 with it.

Section 500.11 requires a covered entity to adopt its own policies and contractual protections for a third-party service provider, including contractual notice to the covered entity of a cybersecurity event, so Part 500 reaches a third-party service provider only through the covered entity's own duties and never binds the provider directly unless the provider independently qualifies as its own covered entity.

Covered online platforms, including social media, gaming, and digital messaging services, must apply privacy-protective default settings for users they know are minors, including restricting contact and profile recommendations from unknown adults, limiting financial transactions, requiring parental approvals for weaker settings, and turning AI companion features off by default.

The act relies on commercially reasonable age assurance, with standards to be set through Attorney General rulemaking, and was enacted in May 2026 as part Y of the state's FY2027 transportation and economic development budget bill.

Operators of websites, online services, apps, and connected products may not process, sell, or share the personal data of a user under 18 for targeted advertising, profiling, or other non-essential purposes without informed consent, and must honor browser or device signals indicating a user is a minor. The Attorney General has said it will exercise enforcement discretion for good-faith compliance efforts while final implementing rules remain pending.

Social media platforms must obtain parental consent before providing an addictive, algorithmically personalized feed to a user they know is a minor, and may not send notifications to minors between midnight and 6 a.m. without parental consent.

The act does not take effect until 180 days after the Attorney General finalizes rules identifying acceptable age-determination and parental-consent methods; a notice of proposed rulemaking was published September 15, 2025, the public comment period closed December 1, 2025, and final rules had not been adopted as of mid-2026.