Regulon — NY-AB-8884 · NY AB 8884 (New York AI Act)

WHAT THIS BILL REGULATES · 6 REQUIREMENT TYPES

How Is This Bill Enforced

Enforcement Authority

Dual enforcement. The attorney general may seek injunctive relief, restitution, and civil penalties in supreme court upon ten days' notice. Private right of action by plenary proceeding for any person harmed by a violation. At the motion-to-dismiss stage, the court presumes the AI system was created or operated in violation and that the violation caused the alleged harm; the defendant must rebut by clear and convincing evidence. A developer may rebut the presumption by demonstrating compliance with the safe harbor under § 89-b.

Private Right of Action

Private right of action by plenary proceeding for any person harmed by a violation.

Penalties

AG enforcement: civil penalty up to $20,000 per violation; injunctive relief and restitution without requiring proof of actual injury. Private right of action: compensatory damages and legal fees to the prevailing party. Court presumes violation and causation at the motion-to-dismiss stage; defendant must rebut by clear and convincing evidence.

Verbatim statutory text on the left; plain-language analysis and a per-section checklist on the right. Numbered markers cross-link to the matching checklist row.

Statutory Text

Analysis & Obligations

Civ. Rights Law § 85

Definitions

§ 85 The following terms shall have the following meanings: 1. "Algorithmic discrimination" means any condition in which the use of an AI system contributes to unjustified differential treatment or impacts, disfavoring people based on their actual or perceived age, race, ethnicity, creed, religion, color, national origin, citizenship or immigration status, sexual orientation, gender identity, gender expression, military status, sex, disability, predisposing genetic characteristics, familial status, marital status, pregnancy, pregnancy outcomes, disability, height, weight, reproductive health care or autonomy, status as a victim of domestic violence or other classification protected under state or federal laws. Algorithmic discrimination shall not include: (a) a developer's or deployer's testing of their own AI system to identify, mitigate, and prevent discriminatory bias; (b) expanding an applicant, customer, or participant pool to increase diversity or redress historical discrimination; or (c) an act or omission by or on behalf of a private club or other establishment that is not in fact open to the public, as set forth in Title II of the federal Civil Rights Act of 1964, 42 U.S.C. section 2000a(e), as amended. 2. "Artificial intelligence system" or "AI system" means a machine-based system or combination of systems, that for explicit and implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Artificial intelligence system shall not include: (a) any system that (i) is used by a business entity solely for internal purposes and (ii) is not used as a substantial factor in a consequential decision; or (b) any software used primarily for basic computerized processes, such as anti-malware, anti-virus, auto-correct functions, calculators, databases, data storage, electronic communications, firewall, internet domain registration, internet website loading, networking, spam and robocall-filtering, spellcheck tools, spreadsheets, web caching, web hosting, or any tool that relates only to internal management affairs such as ordering office supplies or processing payments, and that do not materially affect the rights, liberties, benefits, safety or welfare of any individual within the state. 3. "Auditor" shall refer to an independent entity including but not limited to an individual, non-profit, firm, corporation, partnership, cooperative, association, academic institution, or group affiliated with an academic institution, commissioned to perform an audit. 4. "Consequential decision" means a decision or judgment that has a material, legal or similarly significant effect on an individual's access to, or the cost, terms, or availability of, any of the following: (a) Employment, workers' management, or self-employment, including, but not limited to, all of the following: (i) Pay or promotion; and (ii) Hiring or termination. (b) Education and vocational training, including, but not limited to, all of the following: (i) Accreditation; (ii) Certification; (iii) Admissions; and (iv) Financial aid or scholarships. (c) Housing or lodging, including rental or short-term housing or lodging. (d) Family planning, including adoption services or reproductive services, as well as assessments related to child protective services. (e) Health care or health insurance, including mental health care, dental, or vision. (f) Financial services, including a financial service provided by a mortgage company, mortgage broker, or creditor. (g) Law enforcement activities, including the allocation of law enforcement personnel or assets, the enforcement of laws, maintaining public order, or managing public safety. (h) Legal services. 5. "Deployer" means any person, partnership, association or corporation that offers or uses an AI system for commerce in the state of New York, or provides an AI system for use by the general public in the state of New York. A deployer shall not include any natural person using an AI system for personal use. A developer may also be considered a deployer if its actions satisfy this definition. 6. "Developer" means a person, partnership, or corporation that designs, codes, or produces an AI system, or creates a substantial change with respect to an AI system, whether for its own use in the state of New York or for use by a third party in the state of New York. A deployer may also be considered a developer if its actions satisfy this definition. 7. "Employee" means an individual who performs services for and under the control and direction of an employer for wages or other remuneration, including former employees, or natural persons employed as independent contractors to carry out work in furtherance of an employer's business enterprise who are not themselves employers. 8. "Employer" means any person, firm, partnership, institution, corporation, or association that employs one or more employees. 9. "End user" means any individual or group of individuals that: (a) is the subject of a consequential decision made entirely by or with the assistance of an AI system; or (b) interacts, directly or indirectly, with the relevant AI system on behalf of an individual or group that is the subject of a consequential decision made entirely by or with the assistance of an AI system. 10. "High-risk AI system" means any AI system that, when deployed: (a) is a substantial factor in making a consequential decision; or (b) will have a material impact on the statutory or constitutional rights, civil liberties, safety, or welfare of an individual in the state. 11. "Risk management policy and program" means the risk management policy and program created pursuant to section eighty-nine of this article. 12. "Substantial change" means any new version, new release, or any other update to an AI system that results in significant changes to such AI system's appropriate use cases, key functionality, or expected outcomes. 13. "Substantial factor" means a factor that is (a) material in making a consequential decision, or (b) is capable of altering the outcome of a consequential decision.

Section 85 establishes the foundational definitions for the New York AI Act. Key defined terms include algorithmic discrimination (which covers a broad list of protected characteristics and includes three carve-outs for self-testing, diversity expansion, and private clubs), AI system (with exclusions for purely internal systems and basic software tools), consequential decision (covering eight domains: employment, education, housing, family planning, healthcare, financial services, law enforcement, and legal services), high-risk AI system (any AI system that is a substantial factor in a consequential decision or has material impact on rights, liberties, safety, or welfare), and deployer and developer (with express cross-over provisions allowing each to be treated as the other if their actions so qualify).

Civ. Rights Law § 86

Unlawful discriminatory practices — reasonable care and pre-deployment audit

DeployerDeveloper

(1) 1 A developer or deployer shall take reasonable care to prevent foreseeable risk of algorithmic discrimination that is a consequence of the use, sale, or sharing of a high-risk AI system or a product featuring a high-risk AI system.

(2) 2 Any developer or deployer that uses, sells, or shares a high-risk AI system shall have completed an independent audit, pursuant to section eighty-seven of this article, confirming that the developer or deployer has taken reasonable care to prevent foreseeable risk of algorithmic discrimination with respect to such high-risk AI system.

Section 86 establishes the bill's core anti-discrimination mandate: developers and deployers must take reasonable care to prevent foreseeable algorithmic discrimination arising from their high-risk AI systems. Subdivision 2 conditions the lawful use, sale, or sharing of a high-risk AI system on having completed an independent audit under § 87 confirming that reasonable care was taken. Together, these provisions make both the substantive standard (reasonable care) and the procedural prerequisite (audit completion) elements of the unlawful-discriminatory-practice cause of action.

Compliance actions 2 items

Developers and deployers must take reasonable care to prevent foreseeable risk of algorithmic discrimination arising from the use, sale, or sharing of a high-risk AI system or a product featuring a high-risk AI system.

H-02.1

Developers and deployers must complete an independent audit under § 87 confirming that reasonable care has been taken to prevent foreseeable risk of algorithmic discrimination before using, selling, or sharing a high-risk AI system.

H-02.6

Civ. Rights Law § 86-a

Deployer and developer obligations — notice, opt-out, appeal, and accountability

DeployerDeveloper

(1)(a)(i)–(ii), (1)(b)–(d) 3 Any deployer that employs a high-risk AI system for a consequential decision shall comply with the following requirements; provided, however, that where there is an urgent necessity for a decision to be made to confer a benefit to the end user, including, but not limited to, social benefits, housing access, or dispensing of emergency funds, and compliance with this section would cause imminent detriment to the welfare of the end user, such obligation shall be considered waived; provided further, that nothing in this section shall be construed to waive a natural person's option to request human review of the decision: (i) inform the end user at least five business days prior to the use of such system for the making of a consequential decision in clear, conspicuous, and consumer-friendly terms, made available in each of the languages in which the company offers its end services, that AI systems will be used to make a decision or to assist in making a decision; and (ii) allow sufficient time and opportunity in a clear, conspicuous, and consumer-friendly manner for the consumer to opt-out of the automated consequential decision process and for the decision to be made by a human representative. A consumer may not be punished or face any other adverse action for opting out of a decision by an AI system and the deployer shall render a decision to the consumer within forty-five days. (b) If a deployer employs a high-risk AI system for a consequential decision to determine whether to or on what terms to confer a benefit on an end user, the deployer shall offer the end user the option to waive their right to advance notice of five business days under this subdivision. (c) If the end user clearly and affirmatively waives their right to five business days' notice, the deployer shall then inform the end user as early as practicable before the making of the consequential decision in clear, conspicuous, and consumer-friendly terms, made available in each of the languages in which the company offers its end services, that AI systems will be used to make a decision or to assist in making a decision. The deployer shall allow sufficient time and opportunity in a clear, conspicuous, and consumer-friendly manner for the consumer to opt-out of the automated process and for the decision to be made by a human representative. A consumer may not be punished or face any other adverse action for opting out of a decision by an AI system and the deployer shall render a decision to the consumer within forty-five days. (d) An end user shall be entitled to no more than one opt-out with respect to the same consequential decision within a six-month period.

(2)(a)–(b) 4 Any deployer that employs a high-risk AI system for a consequential decision shall inform the end user within five days in a clear, conspicuous and consumer-friendly manner if a high-risk AI system has been used to make a consequential decision. The deployer shall then provide and explain a process for the end user to appeal the decision, which shall at minimum allow the end user to (i) formally contest the decision, (ii) provide information to support their position, and (iii) obtain meaningful human review of the decision. A deployer shall respond to an end user's appeal within forty-five days of receipt of the appeal. That period may be extended once by forty-five additional days where reasonably necessary, taking into account the complexity and number of appeals. The deployer shall inform the end user of any such extension within forty-five days of receipt of the appeal, together with the reasons for the delay. (b) An end user shall be entitled to no more than one appeal with respect to the same consequential decision in a six-month period.

(3) 5 The deployer or developer of a high-risk AI system is legally responsible for quality and accuracy of all consequential decisions made, including any bias or algorithmic discrimination resulting from the operation of the AI system on their behalf.

(4)–(5) 4. The rights and obligations under this section may not be waived by any person, partnership, association or corporation. 5. With respect to a single consequential decision, an end user may not exercise both its right to opt-out of a consequential decision under subdivision one of this section and its right to appeal a consequential decision under subdivision two of this section.

Section 86-a imposes the bill's most operationally intensive deployer-facing obligations. Subdivision 1 requires advance notice at least five business days before a high-risk AI system makes or assists in a consequential decision, plus an opt-out right allowing the end user to have a human make the decision instead. An emergency waiver applies where compliance would cause imminent detriment to the end user, but even then the right to request human review is preserved. End users may waive the five-day notice for benefit-conferring decisions, but must still receive notice as early as practicable.

Subdivision 2 requires post-decision notice within five days and a formal appeal process with human review, a 45-day response deadline (extendable once), and a cap of one appeal per consequential decision per six-month period. Subdivision 3 establishes that developers and deployers are legally responsible for the quality and accuracy of all consequential decisions, including any bias or algorithmic discrimination. Subdivisions 4 and 5 make the rights non-waivable and prevent end users from exercising both opt-out and appeal with respect to the same decision.

Compliance actions 3 items

Deployers must inform end users at least five business days before a high-risk AI system is used to make or assist in a consequential decision, in clear, conspicuous, and consumer-friendly terms available in each language in which the company offers services. Deployers must provide end users with sufficient time and opportunity to opt out of the automated decision process and have the decision made by a human representative. End users may not be punished or face adverse action for opting out, and the deployer must render a human decision within 45 days. For benefit-conferring decisions, the deployer must offer the end user the option to waive the five-day notice period; if waived, notice must be provided as early as practicable. An emergency waiver applies where compliance would cause imminent detriment to the end user, but the right to request human review is not waived. End users are entitled to no more than one opt-out per consequential decision within a six-month period.

H-01.3

Deployers must inform end users within five days after a high-risk AI system has been used to make a consequential decision. The deployer must provide and explain an appeal process that at minimum allows the end user to (1) formally contest the decision, (2) provide information to support their position, and (3) obtain meaningful human review of the decision. The deployer must respond to an appeal within 45 days, extendable once by 45 additional days where reasonably necessary, with written notice of the extension and reasons. End users are entitled to no more than one appeal per consequential decision within a six-month period.

H-01.4

Developers and deployers of high-risk AI systems are legally responsible for the quality and accuracy of all consequential decisions made by or with the assistance of the AI system, including any bias or algorithmic discrimination resulting from the system's operation on their behalf.

H-01

Civ. Rights Law § 86-b

Whistleblower protections

DeployerDeveloper

(1)(a)–(b) 6 Developers and/or deployers of high-risk AI systems shall not: (a) prevent any of their employees from disclosing information to the attorney general, including through terms and conditions of employment or seeking to enforce terms and conditions of employment, if the employee has reasonable cause to believe the information indicates a violation of this article; or (b) retaliate against an employee for disclosing information to the attorney general pursuant to this section.

(2) An employee harmed by a violation of this article may petition a court for appropriate relief as provided in subdivision five of section seven hundred forty of the labor law.

(3) 7 Developers and deployers of high-risk AI systems shall provide a clear notice to all of their employees working on such AI systems of their rights and responsibilities under this article, including the right of employees of contractors and subcontractors to use the developer's internal process for making protected disclosures pursuant to subdivision four of this section. A developer or deployer is presumed to be in compliance with the requirements of this subdivision if the developer or deployer does either of the following: (a) at all times post and display within all workplaces maintained by the developer or deployer a notice to all employees of their rights and responsibilities under this article, ensure that all new employees receive equivalent notice, and ensure that employees who work remotely periodically receive an equivalent notice; or (b) no less frequently than once every year, provide written notice to all employees of their rights and responsibilities under this article and ensure that the notice is received and acknowledged by all of those employees.

(4) 8 Each developer and deployer shall provide a reasonable internal process through which an employee may anonymously disclose information to the developer or deployer if the employee believes in good faith that the information indicates that the developer or deployer has violated any provision of this article or any other law, or has made false or materially misleading statements related to its risk management policy and program, or failed to disclose known risks to employees, including, at a minimum, a monthly update to the person who made the disclosure regarding the status of the developer's or deployer's investigation of the disclosure and the actions taken by the developer or deployer in response to the disclosure.

(5) This section does not limit protections provided to employees under section seven hundred forty of the labor law.

Section 86-b creates whistleblower protections for employees of developers and deployers of high-risk AI systems. Subdivision 1 prohibits employers from preventing employees from disclosing information to the attorney general where the employee has reasonable cause to believe a violation occurred, and prohibits retaliation for such disclosures. Subdivision 2 provides harmed employees a judicial remedy via the Labor Law § 740 framework. Subdivision 3 requires notice to employees of their rights, with a compliance safe harbor for either continuous workplace posting or annual written distribution. Subdivision 4 mandates an anonymous internal disclosure process with monthly status updates. Subdivision 5 is a savings clause preserving Labor Law § 740 protections.

Compliance actions 3 items

Developers and deployers of high-risk AI systems must not prevent employees from disclosing information to the attorney general — including through employment terms and conditions or enforcement thereof — where the employee has reasonable cause to believe the information indicates a violation of this article. Developers and deployers must not retaliate against employees for making such disclosures.

G-03.3

Developers and deployers must provide clear notice to all employees working on high-risk AI systems of their rights and responsibilities under this article, including contractors' and subcontractors' employees' right to use the internal anonymous disclosure process. Compliance is presumed if the developer or deployer either (1) continuously posts workplace notices, provides equivalent notice to new employees, and periodically provides equivalent notice to remote employees, or (2) annually provides written notice to all employees and ensures receipt and acknowledgment.

G-03.4

Developers and deployers must provide a reasonable internal process through which employees may anonymously disclose information indicating a violation of this article or any other law, false or materially misleading statements related to the risk management policy and program, or failure to disclose known risks. The process must include, at minimum, monthly status updates to the disclosing employee regarding the investigation and actions taken in response.

G-03.1

Civ. Rights Law § 87

Audits

DeployerDeveloper

(1)(a)–(b) 9 Developers of high-risk AI systems shall cause to be conducted third-party audits in accordance with this section. (a) A developer of a high-risk AI system shall complete at least: (i) a first audit within six months after completion of development of the high-risk AI system and the initial offering of the high-risk AI system to a deployer for deployment or, if the developer is first deployer to deploy the high-risk AI system, after initial deployment; and (ii) one audit every one year following the submission of the first audit. (b) A developer audit under this section shall include: (i) an evaluation and determination of whether the developer has taken reasonable care to prevent foreseeable risk of algorithmic discrimination with respect to such high-risk AI system; and (ii) an evaluation of the developer's documented risk management policy and program required under section eighty-nine of this article for conformity with subdivision one of such section eighty-nine.

(2)(a)–(b) 10 Deployers of high-risk AI systems shall cause to be conducted third-party audits in accordance with this section. (a) A deployer of a high-risk AI system shall complete at least: (i) a first audit within six months after initial deployment; (ii) a second audit within one year following the submission of the first audit; and (iii) one audit every two years following the submission of the second audit. (b) A deployer audit under this section shall include: (i) an evaluation and determination of whether the deployer has taken reasonable care to prevent foreseeable risk of algorithmic discrimination with respect to such high-risk AI system; (ii) an evaluation of system accuracy and reliability with respect to such high-risk AI system's deployer-intended and actual use cases; and (iii) an evaluation of the deployer's documented risk management policy and program required under section eighty-nine of this article for conformity with subdivision one of such section eighty-nine.

(3)–(9) 9 3. A deployer or developer may hire more than one auditor to fulfill the requirements of this section. 4. At the attorney general's discretion, the attorney general may: (a) promulgate further rules as necessary to ensure that audits under this section assess whether or not AI systems produce algorithmic discrimination and otherwise comply with the provisions of this article; and (b) recommend an updated AI system auditing framework to the legislature, where such recommendations are based on a standard or framework (i) designed to evaluate the risks of AI systems, and (ii) that is nationally or internationally recognized and consensus-driven, including but not limited to a relevant framework or standard created by the International Standards Organization. 5. The independent auditor shall have complete and unredacted copies of all reports previously filed by the deployer or developer under section eighty-eight of this article. 6. An audit conducted under this section may be completed in part, but shall not be completed entirely, with the assistance of an AI system. (a) Acceptable auditor uses of an AI system include, but are not limited to: (i) use of an audited high-risk AI system in a controlled environment without impacts on end users for system testing purposes; or (ii) detecting patterns in the behavior of an audited AI system. (b) An auditor shall not: (i) use a different high-risk AI system that is not the subject of an audit to complete an audit; or (ii) use an AI system to draft an audit under this section without meaningful human review and oversight. 7. (a) An auditor shall be an independent entity including but not limited to an individual, non-profit, firm, corporation, partnership, cooperative, or association. (b) For the purposes of this article, no auditor may be commissioned by a developer or deployer of a high-risk AI system if such entity: (i) has already been commissioned to provide any auditing or non-auditing service, including but not limited to financial auditing, cybersecurity auditing, or consulting services of any type, to the commissioning company in the past twelve months; or (ii) is, will be, or plans to be engaged in the business of developing or deploying an AI system that can compete commercially with such developer's or deployer's high-risk AI system in the five years following an audit. (c) Fees paid to auditors may not be contingent on the result of the audit and the commissioning company shall not provide any incentives or bonuses for a positive audit result. 8. The attorney general may promulgate further rules to ensure (a) the independence of auditors under this section, and (b) that teams conducting audits incorporate feedback from communities that may foreseeably be the subject of algorithmic discrimination with respect to the AI system being audited. 9. If a developer or deployer has an audit completed for the purpose of complying with another applicable federal, state, or local law or regulation, and the audit otherwise satisfies all other requirements of this section, such audit shall be deemed to satisfy the requirements of this section.

Section 87 establishes the independent third-party audit regime for high-risk AI systems. Developers must complete a first audit within six months of initial offering or deployment, then annually thereafter. Deployers must complete a first audit within six months of initial deployment, a second within one year, then biennially. Developer audits evaluate (1) reasonable care against algorithmic discrimination and (2) the risk management policy and program. Deployer audits add (3) system accuracy and reliability for intended and actual use cases.

The section imposes strict auditor independence requirements: no prior commercial relationship within twelve months, no competitive AI development plans within five years, non-contingent fees, and no incentives for positive results. Audits may use AI partially (e.g., controlled-environment testing) but may not be completed entirely by AI. A cross-compliance safe harbor allows audits completed for other applicable laws to satisfy these requirements if they meet all other conditions. The attorney general may promulgate further rules on audit methodology and auditor independence.

Compliance actions 2 items

Developers must cause independent third-party audits to be conducted: a first audit within six months of initial offering or deployment, then annually. Each developer audit must evaluate (1) whether the developer has taken reasonable care to prevent foreseeable algorithmic discrimination and (2) the developer's risk management policy and program for conformity with § 89. The auditor must be independent — no prior commercial relationship with the commissioning company in the past 12 months, no competitive AI business plans for five years post-audit, non-contingent fees, and no incentives for positive results. Audits may use AI tools partially (e.g., controlled-environment testing or pattern detection) but may not be completed entirely by AI and may not be drafted by AI without meaningful human review. The auditor must receive complete and unredacted copies of all prior § 88 reports. An audit completed for another applicable federal, state, or local law satisfies these requirements if it meets all other conditions of this section.

H-02.6

Deployers must cause independent third-party audits to be conducted: a first audit within six months of initial deployment, a second within one year of the first, then biennially. Each deployer audit must evaluate (1) whether the deployer has taken reasonable care to prevent foreseeable algorithmic discrimination, (2) system accuracy and reliability for the deployer's intended and actual use cases, and (3) the deployer's risk management policy and program for conformity with § 89. The same auditor independence requirements, AI-use restrictions, and cross-compliance safe harbor that apply to developer audits under this section apply to deployer audits.

H-02.6

Civ. Rights Law § 88

High-risk AI system reporting requirements

DeployerDeveloper

(1)–(2) 11 1. Every developer and deployer of a high-risk AI system shall comply with the reporting requirements of this section. 2. Together with each report required to be filed under this section, every developer and deployer shall file with the attorney general a copy of the last completed independent audit required by this article.

(3)(a)–(b) 12 3. Developers of high-risk AI systems shall complete and file with the attorney general reports in accordance with this subdivision. (a) A developer of a high-risk AI system shall complete and file with the attorney general at least: (i) a first report within six months after completion of development of the high-risk AI system and the initial offering of the high-risk AI system to a deployer for deployment or, if the developer is first deployer to deploy the high-risk AI system, after initial deployment; (ii) one report annually following the submission of the first report; and (iii) one report within six months of any substantial change to the high-risk AI system. (b) A developer report under this section shall include: (i) a description of the system including: (A) the uses of the high-risk AI system that the developer intends; and (B) any explicitly unintended or disallowed uses of the high-risk AI system; (ii) an overview of how the high-risk AI system was developed; (iii) an overview of the high-risk AI system's training data; and (iv) any other information necessary to allow a deployer to: (A) understand the outputs and monitor the system for compliance with this article; and (B) fulfill its duties under this article.

(4)(a)–(c) 13 4. Deployers of high-risk AI systems shall complete and file with the attorney general reports in accordance with this subdivision. (a) A deployer of a high-risk AI system shall complete and file with the attorney general at least: (i) a first report within six months after initial deployment; (ii) a second report within one year following the completion and filing of the first report; (iii) one report every two years following the completion and filing of the second report; and (iv) one report within six months of any substantial change to the high-risk AI system. (b) A deployer report under this section shall include: (i) a description of the system including: (A) the deployer's actual, intended, or planned uses of the high-risk AI system with respect to consequential decisions; and (B) whether the deployer is using the high-risk AI system for any developer unintended or disallowed uses; and (ii) an impact assessment including: (A) whether the high-risk AI system poses a risk of algorithmic discrimination and the steps taken to address the risk of algorithmic discrimination; (B) if the high-risk AI system is or will be monetized, how it is or is planned to be monetized; and (C) an evaluation of the costs and benefits to consumers and other end users. (c) A deployer that is also a developer and is required to submit reports under subdivision three of this section may submit a single joint report provided it contains the information required in this subdivision.

(5)(a)–(b) 5. The attorney general shall: (a) promulgate rules for a process whereby developers and deployers may request redaction of portions of reports required under this section to ensure that they are not required to disclose sensitive and protected information; and (b) maintain an online database that is accessible to the general public with reports, redacted in accordance with this subdivision, and audits required by this article, which database shall be updated biannually.

(6)(a)–(b) 14 6. For high-risk AI systems which are already in deployment at the time of the effective date of this article, developers and deployers shall have eighteen months from such effective date to complete and file the first report and associated independent audit required by this article. (a) Each developer of a high-risk AI system shall thereafter file at least one report annually following the submission of the first report under this subdivision. (b) Each deployer of a high-risk AI system shall thereafter file at least one report every two years following the submission of the first report under this subdivision.

Section 88 creates a periodic reporting regime requiring both developers and deployers to file reports with the attorney general. Developer reports are due within six months of initial offering or deployment, then annually, and within six months of any substantial change. Developer reports must describe the system (intended and disallowed uses), development methodology, training data, and information necessary for deployer compliance. Deployer reports follow a similar schedule (six months, one year, then biennially) plus a report within six months of substantial changes, and must include system descriptions plus an impact assessment covering algorithmic discrimination risk, monetization, and cost-benefit analysis.

Each report must be accompanied by the most recent independent audit. The attorney general must maintain a publicly accessible online database of redacted reports and audits, updated biannually. For systems already deployed at the effective date, an 18-month grace period applies for the first report and audit.

Compliance actions 4 items

Developers and deployers must file with the attorney general a copy of the last completed independent audit together with each report filed under § 88.

R-02.1

Developers must file reports with the attorney general on the following schedule: (1) a first report within six months of initial offering or deployment, (2) annually thereafter, and (3) within six months of any substantial change. Each developer report must include: a system description covering intended and disallowed uses, an overview of development methodology, an overview of training data, and any additional information necessary for deployers to understand outputs, monitor compliance, and fulfill their obligations under this article.

R-02.1

Deployers must file reports with the attorney general on the following schedule: (1) a first report within six months of initial deployment, (2) a second report within one year of the first, (3) biennially thereafter, and (4) within six months of any substantial change. Each deployer report must include: a system description covering actual, intended, or planned uses for consequential decisions and whether the system is being used for any developer-disallowed purposes; and an impact assessment covering (A) algorithmic discrimination risk and mitigation steps, (B) monetization details if applicable, and (C) a cost-benefit evaluation for consumers and end users. A deployer that is also a developer may submit a single joint report covering both sets of requirements.

R-02.1

For high-risk AI systems already in deployment at the effective date, developers and deployers have 18 months to complete and file the first report and associated independent audit. Developers must thereafter file annually; deployers must file biennially.

R-02.1

Civ. Rights Law § 89

Risk management policy and program

DeployerDeveloper

(1)(a)–(d) 15 Each developer or deployer of high-risk AI systems shall plan, document, and implement a risk management policy and program to govern development or deployment, as applicable, of such high-risk AI system. The risk management policy and program shall specify and incorporate the principles, processes, and personnel that the deployer uses to identify, document, and mitigate known or reasonably foreseeable risks of algorithmic discrimination covered under subdivision one of section eighty-six of this article. The risk management policy and program shall be an iterative process planned, implemented, and regularly and systematically reviewed and updated over the life cycle of a high-risk AI system, requiring regular, systematic review and updates, including updates to documentation. A risk management policy and program implemented and maintained pursuant to this section shall be reasonable considering: (a) The guidance and standards set forth in: (i) version 1.0 of the "Artificial Intelligence Risk Management Framework" published by the National Institute of Standards and Technology in the United States department of commerce, or (ii) another substantially equivalent framework selected at the discretion of the attorney general, if such framework was designed to manage risks associated with AI systems, is nationally or internationally recognized and consensus-driven, and is at least as stringent as version 1.0 of the "Artificial Intelligence Risk Management Framework" published by the National Institute of Standards and Technology; (b) The size and complexity of the developer or deployer; (c) The nature, scope, and intended uses of the high-risk AI system developed or deployed; and (d) The sensitivity and volume of data processed in connection with the high-risk AI system.

(2)–(3) 15 2. A risk management policy and program implemented pursuant to subdivision one of this section may cover multiple high-risk AI systems developed by the same developer or deployed by the same deployer if sufficient. 3. The attorney general may require a developer or a deployer to disclose the risk management policy and program implemented pursuant to subdivision one of this section in a form and manner prescribed by the attorney general. The attorney general may evaluate the risk management policy and program to ensure compliance with this section.

Section 89 requires each developer and deployer of high-risk AI systems to plan, document, and implement a risk management policy and program governing the development or deployment of such systems. The program must identify, document, and mitigate known or reasonably foreseeable risks of algorithmic discrimination, and must be iterative — regularly and systematically reviewed and updated over the system's life cycle. Reasonableness is assessed against the NIST AI RMF v1.0 or an equivalent framework selected by the attorney general, as well as organizational size, system scope, and data sensitivity. A single program may cover multiple systems. The attorney general may require disclosure of the program and evaluate it for compliance.

Compliance actions 1 item

Developers and deployers of high-risk AI systems must plan, document, and implement a risk management policy and program governing the development or deployment of each high-risk AI system. The program must specify the principles, processes, and personnel used to identify, document, and mitigate known or reasonably foreseeable risks of algorithmic discrimination. The program must be iterative — regularly and systematically reviewed and updated over the system's life cycle. Reasonableness is assessed against the NIST AI RMF v1.0 (or a substantially equivalent framework approved by the attorney general), organizational size and complexity, system scope and intended uses, and data sensitivity and volume. A single program may cover multiple high-risk AI systems if sufficient. The attorney general may require disclosure of the program in a prescribed form and evaluate it for compliance.

G-01.1

Civ. Rights Law § 89-a

Social scoring AI systems prohibited

DeployerDeveloper

§ 89-a 16 No person, partnership, association or corporation shall develop, deploy, use, or sell an AI system which evaluates or classifies the trustworthiness of natural persons over a certain period of time based on their social behavior or known or predicted personal or personality characteristics, with the social score leading to any of the following: 1. differential treatment of certain natural persons or whole groups thereof in social contexts which are unrelated to the contexts in which the data was originally generated or collected; 2. differential treatmentof certain natural persons or whole groups thereof that is unjustified or disproportionate to their social behavior or its gravity; or 3. the infringement of any right guaranteed under the United States constitution, the New York constitution, or state or federal law.

Section 89-a categorically prohibits any person, partnership, association, or corporation from developing, deploying, using, or selling a social scoring AI system. The prohibition targets AI systems that evaluate or classify natural persons' trustworthiness over time based on social behavior or personal or personality characteristics, where the resulting score leads to unjustified differential treatment, disproportionate treatment, or infringement of constitutional or statutory rights.

Compliance actions 1 item

No person, partnership, association, or corporation may develop, deploy, use, or sell an AI system that evaluates or classifies the trustworthiness of natural persons over time based on their social behavior or known or predicted personal or personality characteristics, where the resulting social score leads to (1) differential treatment in social contexts unrelated to the original data collection context, (2) differential treatment that is unjustified or disproportionate to the person's social behavior or its gravity, or (3) infringement of any right guaranteed under the U.S. Constitution, New York Constitution, or state or federal law.

S-02.1

Civ. Rights Law § 89-b

Developer safe harbor

Developer

§ 89-b 17 A developer may be exempt from its duties and obligations under sections eighty-six, eighty-six-a, eighty-six-b, eighty-seven, eighty-eight, and eighty-nine of this article if such developer: 1. receives a written and signed contractual agreement from each deployer authorized to use the artificial intelligence system developed by such developer, including the developer if they are also a deployer, that such artificial intelligence system will not be used as a high-risk AI system; 2. implements reasonable technical safeguards designed to prevent or detect high-risk AI system use cases or otherwise demonstrates reasonable steps taken to ensure that any unauthorized deployments of its AI systems are not being used as a high-risk AI system; 3. prominently displays on its website, in marketing materials, and in all licensing agreements offered to prospective deployers of its AI system that the AI system cannot be used as a high-risk AI system; and 4. maintains records of deployer agreements for a period of not less than five years.

Section 89-b creates a safe harbor allowing developers to be exempt from all substantive obligations under §§ 86, 86-a, 86-b, 87, 88, and 89 if four conditions are met: (1) each authorized deployer provides a written, signed contractual agreement that the AI system will not be used as a high-risk AI system; (2) the developer implements reasonable technical safeguards to prevent or detect high-risk use; (3) the developer prominently displays on its website, marketing materials, and licensing agreements that the system cannot be used as a high-risk AI system; and (4) the developer maintains deployer agreement records for at least five years. This safe harbor also functions as a rebuttal mechanism in private litigation under § 89-c(3)(b).

Compliance actions 1 item

A developer may be exempt from all duties under §§ 86, 86-a, 86-b, 87, 88, and 89 if the developer: (1) obtains a written, signed contractual agreement from each authorized deployer (including the developer itself if also deploying) certifying that the AI system will not be used as a high-risk AI system; (2) implements reasonable technical safeguards to prevent or detect high-risk use cases; (3) prominently displays on its website, in marketing materials, and in all licensing agreements that the AI system cannot be used as a high-risk AI system; and (4) maintains records of deployer agreements for at least five years.

G-01.3

Civ. Rights Law § 89-c

Enforcement

(1)–(4) 1. Whenever there shall be a violation of section eighty-six-a, eighty-six-b, eighty-seven, eighty-eight, eighty-nine, or eighty-nine-a of this article, an application may be made by the attorney general in the name of the people of the state of New York, to the supreme court having jurisdiction to issue an injunction, and upon notice to the respondent of not less than ten days, to enjoin and restrain the continuance of such violation; and if it shall appear to the satisfaction of the court that the respondent has, in fact, violated this article, an injunction may be issued by the court, enjoining and restraining any further violations, without requiring proof that any person has, in fact, been injured or damaged thereby. In any such proceeding, the court may make allowances to the attorney general as provided in paragraph six of subdivision (a) of section eighty-three hundred three of the civil practice law and rules, and direct restitution. Whenever the court shall determine that a violation of this article has occurred, the court may impose a civil penalty of not more than twenty thousand dollars for each violation. 2. There shall be a private right of action by plenary proceeding for any person harmed by any violation of section eighty-six-a, eighty-six-b, eighty-seven, eighty-eight, eighty-nine, or eighty-nine-a of this article by any natural person or entity. The court shall award compensatory damages and legal fees to the prevailing party. 3. In evaluating any motion to dismiss a plenary proceeding commenced pursuant to subdivision two of this section, the court shall presume the specified AI system was created and/or operated in violation of a specified law or laws and that such violation caused the harm or harms alleged. (a) A defendant can rebut presumptions made pursuant to this subdivision through clear and convincing evidence that the specified AI system did not cause the harm or harms alleged and/or did not violate the alleged law or laws. An algorithmic audit can be considered as evidence in rebutting such presumptions, but the mere existence of such an audit, without additional evidence, shall not be considered clear and convincing evidence. (b) With respect to a violation of section eighty-six-a, eighty-six-b, eighty-seven, eighty-eight, or eighty-nine of this article, a developer can rebut presumptions made pursuant to this subdivision through clear and convincing evidence that it has complied with the duties under section eighty-nine-b of this article. (c) Where such presumptions are not rebutted pursuant to this subdivision, the action shall not be dismissed. (d) Where such presumptions are rebutted pursuant to this subdivision, a motion to dismiss an action shall be adjudicated without any consideration of this section. 4. The supreme court in the state shall have jurisdiction over any action, claim, or lawsuit to enforce the provisions of this article.

Section 89-c establishes dual enforcement. Subdivision 1 authorizes the attorney general to seek injunctive relief, restitution, and civil penalties up to $20,000 per violation in supreme court, without requiring proof of actual injury. Subdivision 2 creates a private right of action for any person harmed, with compensatory damages and legal fees to the prevailing party. Subdivision 3 creates a plaintiff-favorable evidentiary presumption: at the motion-to-dismiss stage, courts must presume the AI system was operated in violation and that the violation caused the harm alleged. The defendant must rebut by clear and convincing evidence; an audit alone is insufficient. Developers may rebut by showing compliance with the § 89-b safe harbor. Subdivision 4 vests jurisdiction in supreme court.

Civ. Rights Law § 89-d

Severability

§ 89-d If any clause, sentence, paragraph, subdivision, section or part of this article shall be adjudged by any court of competent jurisdiction to be invalid, such judgment shall not affect, impair, or invalidate the remainder thereof, but shall be confined in its operation to the clause, sentence, paragraph, subdivision, section, or part thereof directly involved in the controversy in which such judgment shall have been made.

Standard severability provision preserving the remainder of the article if any clause, sentence, paragraph, subdivision, section, or part is adjudged invalid.

Exec. Law § 296(23)

Unlawful discriminatory practice under Executive Law

DeployerDeveloper

(23) 18 It shall be an unlawful discriminatory practice under this section for a deployer or a developer, as such terms are defined in section eighty-five of the civil rights law, to engage in an unlawful discriminatory practice under section eighty-six of the civil rights law.

Section 4 of the bill amends the Executive Law to add a new subdivision 23 to § 296, making it an unlawful discriminatory practice under the Human Rights Law for a deployer or developer to engage in an unlawful discriminatory practice under Civil Rights Law § 86. This cross-reference integrates the bill's anti-discrimination mandate into the existing state human rights enforcement framework administered by the Division of Human Rights.

Compliance actions 1 item

It is an unlawful discriminatory practice under the New York Human Rights Law (Executive Law § 296) for a developer or deployer to engage in an unlawful discriminatory practice under Civil Rights Law § 86 — i.e., to fail to take reasonable care to prevent foreseeable algorithmic discrimination or to use, sell, or share a high-risk AI system without a completed independent audit.

H-02

Safety Harmful content prevention · self-harm escalation

1 item

S-02.1

Human Oversight Decision review · appeal pathways

8 items

H-02.1

H-02.6

H-01.3

H-01.4

H-01

H-02.6

H-02

Governance Risk management · accountability programs

5 items

G-03.3

G-03.4

G-03.1

G-01.1

G-01.3

Regulatory Reporting · disclosure · audit

4 items

R-02.1

Developers and deployers must file with the attorney general a copy of the last completed independent audit together with each report filed under § 88.

R-02.1

Passage Likelihood

Medium

Status Introduced

Chamber No passage

Committee No action

Majority party Yes

Bipartisan Yes

Prior session None

Legislative History

2025-06-09 referred to science and technology

2025-06-11 reference changed to ways and means

2026-01-07 referred to ways and means

2026-01-12 reference changed to science and technology

Sources

Full Text ↗

Entry Last Reviewed

2026-05-20

AI generated