S 1962: Enacts the "New York artificial intelligence consumer protection act", in relation to preventing the use of artificial intelligence algorithms to discriminate against protected classes.

  
  STATE OF NEW YORK ________________________________________________________________________ 1962 2025-2026 Regular Sessions  IN SENATE January 14, 2025 ___________ Introduced by Sen. GONZALEZ -- read twice and ordered printed, and when printed to be committed to the Committee on Internet and Technology AN ACT to amend the general business law, in relation to preventing the use of artificial intelligence algorithms to discriminate against protected classes The People of the State of New York, represented in Senate and Assem- bly, do enact as follows: 1 Section 1. Short title. This act shall be known and may be cited as 2 the "New York artificial intelligence consumer protection act". 3 § 2. The general business law is amended by adding a new article 45-A 4 to read as follows: 5 ARTICLE 45-A 6 NEW YORK ARTIFICIAL INTELLIGENCE CONSUMER PROTECTION ACT 7 Section 1550. Definitions. 8 1551. Required documentation. 9 1552. Risk management. 10 1553. Technical documentation. 11 1554. Required disclosure. 12 1555. Preemption. 13 1556. Enforcement. 14 § 1550. Definitions. For the purposes of this article, the following 15 terms shall have the following meanings: 16 1. "Algorithmic discrimination": 17 (a) shall mean any condition in which the use of an artificial intel- 18 ligence decision system results in any unlawful differential treatment 19 or impact that disfavors any individual or group of individuals on the 20 basis of their actual or perceived age, color, disability, ethnicity, 21 genetic information, English language proficiency, national origin, 22 race, religion, reproductive health, sex, veteran status, or other clas- 23 sification protected pursuant to state or federal law; and 24 (b) shall not include: EXPLANATION--Matter in italics (underscored) is new; matter in brackets [ ] is old law to be omitted. LBD01361-01-5 
 S. 1962 2 1 (i) the offer, license, or use of a high-risk artificial intelligence 2 decision system by a developer or deployer for the sole purpose of: 3 (A) such developer's or deployer's self-testing to identify, mitigate, 4 or prevent discrimination or otherwise ensure compliance with state and 5 federal law; or 6 (B) expanding an applicant, customer, or participant pool to increase 7 diversity or redress historic discrimination; or 8 (ii) an act or omission by or on behalf of a private club or other 9 establishment not open to the general public, as set forth in title II 10 of the Civil Rights Act of 1964, 42 U.S.C. § 2000a(e), as amended. 11 2. "Artificial intelligence decision system" shall mean any computa- 12 tional process, derived from machine learning, statistical modeling, 13 data analytics, or artificial intelligence, that issues simplified 14 output, including any content, decision, prediction, or recommendation, 15 that is used to substantially assist or replace discretionary decision 16 making for making consequential decisions that impact consumers. 17 3. "Bias and governance audit" means an impartial evaluation by an 18 independent auditor, which shall include, at a minimum, the testing of 19 an artificial intelligence decision system to assess such system's 20 disparate impact on employees because of such employee's age, race, 21 creed, color, ethnicity, national origin, disability, citizenship or 22 immigration status, marital or familial status, military status, reli- 23 gion, or sex, including sexual orientation, gender identity, gender 24 expression, pregnancy, pregnancy outcomes, and reproductive healthcare 25 choices. 26 4. "Consequential decision" shall mean any decision that has a materi- 27 al legal or similarly significant effect on the provision or denial to 28 any consumer of, or the cost or terms of, any: 29 (a) education enrollment or education opportunity; 30 (b) employment or employment opportunity; 31 (c) financial or lending service; 32 (d) essential government service; 33 (e) health care service, as defined in section 42 U.S.C. § 324(d)(2), 34 as amended; 35 (f) housing or housing opportunity; 36 (g) insurance; or 37 (h) legal service. 38 5. "Consumer" shall mean any New York state resident. 39 6. "Deploy" shall mean to use a high-risk artificial intelligence 40 decision system. 41 7. "Deployer" shall mean any person doing business in this state that 42 deploys a high-risk artificial intelligence decision system. 43 8. "Developer" shall mean any person doing business in this state that 44 develops, or intentionally and substantially modifies, an artificial 45 intelligence decision system. 46 9. "General-purpose artificial intelligence model": 47 (a) shall mean any form of artificial intelligence decision system 48 that: 49 (i) displays significant generality; 50 (ii) is capable of competently performing a wide range of distinct 51 tasks; and 52 (iii) can be integrated into a variety of downstream applications or 53 systems; and 54 (b) shall not include any artificial intelligence model that is used 55 for development, prototyping, and research activities before such arti- 56 ficial intelligence model is released on the market. 
 S. 1962 3 1 10. "High-risk artificial intelligence decision system": 2 (a) shall mean any artificial intelligence decision system that, when 3 deployed, makes, or is a substantial factor in making, a consequential 4 decision; and 5 (b) shall not include: 6 (i) any artificial intelligence decision system that is intended to: 7 (A) perform any narrow procedural task; or 8 (B) detect decision-making patterns, or deviations from decision-mak- 9 ing patterns, unless such artificial intelligence decision system is 10 intended to replace or influence any assessment previously completed by 11 an individual without sufficient human review; or 12 (ii) unless the technology, when deployed, makes, or is a substantial 13 factor in making, a consequential decision: 14 (A) any anti-fraud technology that does not make use of facial recog- 15 nition technology; 16 (B) any artificial intelligence-enabled video game technology; 17 (C) any anti-malware, anti-virus, calculator, cybersecurity, database, 18 data storage, firewall, Internet domain registration, Internet-web-site 19 loading, networking, robocall-filtering, spam-filtering, spellchecking, 20 spreadsheet, web-caching, web-hosting, or similar technology; 21 (D) any technology that performs tasks exclusively related to an enti- 22 ty's internal management affairs, including, but not limited to, order- 23 ing office supplies or processing payments; or 24 (E) any technology that communicates with consumers in natural 25 language for the purpose of providing consumers with information, making 26 referrals or recommendations, and answering questions, and is subject to 27 an accepted use policy that prohibits generating content that is discri- 28 minatory or harmful. 29 11. "Intentional and substantial modification": 30 (a) shall mean any deliberate change made to: 31 (i) an artificial intelligence decision system that results in any new 32 reasonably foreseeable risk of algorithmic discrimination; or 33 (ii) a general-purpose artificial intelligence model that: 34 (A) affects compliance of the general-purpose artificial intelligence 35 model; 36 (B) materially changes the purpose of the general-purpose artificial 37 intelligence model; or 38 (C) results in any new reasonably foreseeable risk of algorithmic 39 discrimination; and 40 (b) shall not include any change made to a high-risk artificial intel- 41 ligence decision system, or the performance of a high-risk artificial 42 intelligence decision system, if: 43 (i) the high-risk artificial intelligence decision system continues to 44 learn after such high-risk artificial intelligence decision system is: 45 (A) offered, sold, leased, licensed, given or otherwise made available 46 to a deployer; or 47 (B) deployed; and 48 (ii) such change: 49 (A) is made to such high-risk artificial intelligence decision system 50 as a result of any learning described in subparagraph (i) of this para- 51 graph; 52 (B) was predetermined by the deployer, or the third party contracted 53 by the deployer, when such deployer or third party completed the initial 54 impact assessment of such high-risk artificial intelligence decision 55 system pursuant to subdivision three of section one thousand five 56 hundred fifty-two of this article; and 
 S. 1962 4 1 (C) is included in the technical documentation for such high-risk 2 artificial intelligence decision system. 3 12. "Person" shall mean any individual, association, corporation, 4 limited liability company, partnership, trust or other legal entity 5 authorized to do business in this state. 6 13. "Red-teaming" shall mean an exercise that is conducted to identify 7 the potential adverse behaviors or outcomes of an artificial intelli- 8 gence decision system and how such behaviors or outcomes occur, and 9 stress test the safeguards against such adverse behaviors or outcomes. 10 14. "Substantial factor": 11 (a) shall mean a factor that: 12 (i) assists in making a consequential decision; 13 (ii) is capable of altering the outcome of a consequential decision; 14 and 15 (iii) is generated by an artificial intelligence decision system; and 16 (b) includes, but is not limited to, any use of an artificial intelli- 17 gence decision system to generate any content, decision, prediction, or 18 recommendation concerning a consumer that is used as a basis to make a 19 consequential decision concerning such consumer. 20 15. "Synthetic digital content" shall mean any digital content, 21 including, but not limited to, any audio, image, text, or video, that is 22 produced or manipulated by an artificial intelligence decision system, 23 including, but not limited to, a general-purpose artificial intelligence 24 model. 25 16. "Trade secret" shall mean any form and type of financial, busi- 26 ness, scientific, technical, economic, or engineering information, 27 including, but not limited to, a pattern, plan, compilation, program 28 device, formula, design, prototype, method, technique, process, proce- 29 dure, program, or code, whether tangible or intangible, and whether 30 stored, compiled, or memorialized physically, electronically, graph- 31 ically, photographically, or in writing, that: 32 (a) derives independent economic value, whether actual or potential, 33 from not being generally known to, or readily ascertainable by proper 34 means by, other persons who can obtain economic value from its disclo- 35 sure or use; and 36 (b) is the subject of efforts that are reasonable under the circum- 37 stances to maintain its secrecy. 38 § 1551. Required documentation. 1. (a) Beginning on January first, two 39 thousand twenty-seven, each developer of a high-risk artificial intelli- 40 gence decision system shall use reasonable care to protect consumers 41 from any known or reasonably foreseeable risks of algorithmic discrimi- 42 nation arising from the intended and contracted uses of a high-risk 43 artificial intelligence decision system. In any enforcement action 44 brought on or after such date by the attorney general pursuant to this 45 article, there shall be a rebuttable presumption that a developer used 46 reasonable care as required pursuant to this subdivision if: 47 (i) the developer complied with the provisions of this section; and 48 (ii) an independent third party identified by the attorney general 49 pursuant to paragraph (b) of this subdivision and retained by the devel- 50 oper completed bias and governance audits for the high-risk artificial 51 intelligence decision system. 52 (b) No later than January first, two thousand twenty-six, and at least 53 annually thereafter, the attorney general shall: 54 (i) identify independent third parties who, in the attorney general's 55 opinion, are qualified to complete bias and governance audits for the 56 purposes of subparagraph (ii) of paragraph (a) of this subdivision; and 
 S. 1962 5 1 (ii) publish a list of such independent third parties available on the 2 attorney general's website. 3 2. Beginning on January first, two thousand twenty-seven, and except 4 as provided in subdivision five of this section, a developer of a high- 5 risk artificial intelligence decision system shall make available to 6 each deployer or other developer the following information: 7 (a) A general statement describing the reasonably foreseeable uses, 8 and the known harmful or inappropriate uses, of such high-risk artifi- 9 cial intelligence decision system; 10 (b) Documentation disclosing: 11 (i) high-level summaries of the type of data used to train such high- 12 risk artificial intelligence decision system; 13 (ii) the known or reasonably foreseeable limitations of such high-risk 14 artificial intelligence decision system, including, but not limited to, 15 the known or reasonably foreseeable risks of algorithmic discrimination 16 arising from the intended uses of such high-risk artificial intelligence 17 decision system; 18 (iii) the purpose of such high-risk artificial intelligence decision 19 system; 20 (iv) the intended benefits and uses of such high-risk artificial 21 intelligence decision system; and 22 (v) any other information necessary to enable such deployer or other 23 developer to comply with the provisions of this article; 24 (c) Documentation describing: 25 (i) how such high-risk artificial intelligence decision system was 26 evaluated for performance, and mitigation of algorithmic discrimination, 27 before such high-risk artificial intelligence decision system was 28 offered, sold, leased, licensed, given, or otherwise made available to 29 such deployer or other developer; 30 (ii) the data governance measures used to cover the training datasets 31 and examine the suitability of data sources, possible biases, and appro- 32 priate mitigation; 33 (iii) the intended outputs of such high-risk artificial intelligence 34 decision system; 35 (iv) the measures such deployer or other developer has taken to miti- 36 gate any known or reasonably foreseeable risks of algorithmic discrimi- 37 nation that may arise from deployment of such high-risk artificial 38 intelligence decision system; and 39 (v) how such high-risk artificial intelligence decision system should 40 be used, not be used, and be monitored by an individual when such high- 41 risk artificial intelligence decision system is used to make, or as a 42 substantial factor in making, a consequential decision; and 43 (d) Any additional documentation that is reasonably necessary to 44 assist a deployer or other developer to: 45 (i) understand the outputs of such high-risk artificial intelligence 46 decision system; and 47 (ii) monitor the performance of such high-risk artificial intelligence 48 decision system for risks of algorithmic discrimination. 49 3. (a) Except as provided in subdivision five of this section, any 50 developer that, on or after January first, two thousand twenty-seven, 51 offers, sells, leases, licenses, gives, or otherwise makes available to 52 a deployer or other developer a high-risk artificial intelligence deci- 53 sion system shall, to the extent feasible, make available to such 54 deployers and other developers the documentation and information relat- 55 ing to such high-risk artificial intelligence decision system necessary 56 for a deployer, or the third party contracted by a deployer, to complete 
 S. 1962 6 1 an impact assessment pursuant to this article. The developer shall make 2 such documentation and information available through artifacts such as 3 model cards, dataset cards, or other impact assessments. 4 (b) A developer that also serves as a deployer for any high-risk arti- 5 ficial intelligence decision system shall not be required to generate 6 the documentation and information required pursuant to this section 7 unless such high-risk artificial intelligence decision system is 8 provided to an unaffiliated entity acting as a deployer. 9 4. (a) Beginning on January first, two thousand twenty-seven, each 10 developer shall publish, in a manner that is clear and readily avail- 11 able, on such developer's website, or a public use case inventory, a 12 statement summarizing: 13 (i) the types of high-risk artificial intelligence decision systems 14 that such developer: 15 (A) has developed or intentionally and substantially modified; and 16 (B) currently makes available to a deployer or other developer; and 17 (ii) how such developer manages any known or reasonably foreseeable 18 risks of algorithmic discrimination that may arise from the development 19 or intentional and substantial modification of the types of high-risk 20 artificial intelligence decision systems described in subparagraph (i) 21 of this subdivision. 22 (b) Each developer shall update the statement described in paragraph 23 (a) of this subdivision: 24 (i) as necessary to ensure that such statement remains accurate; and 25 (ii) no later than ninety days after the developer intentionally and 26 substantially modifies any high-risk artificial intelligence decision 27 system described in subparagraph (i) of paragraph (a) of this subdivi- 28 sion. 29 5. Nothing in subdivisions two or four of this section shall be 30 construed to require a developer to disclose any information: 31 (a) that is a trade secret or otherwise protected from disclosure 32 pursuant to state or federal law; or 33 (b) the disclosure of which would present a security risk to such 34 developer. 35 6. Beginning on January first, two thousand twenty-seven, the attorney 36 general may require that a developer disclose to the attorney general, 37 as part of an investigation conducted by the attorney general and in a 38 form and manner prescribed by the attorney general, the general state- 39 ment or documentation described in subdivision two of this section. The 40 attorney general may evaluate such general statement or documentation to 41 ensure compliance with the provisions of this section. In disclosing 42 such general statement or documentation to the attorney general pursuant 43 to this subdivision, the developer may designate such general statement 44 or documentation as including any information that is exempt from 45 disclosure pursuant to subdivision five of this section or article six 46 of the public officers law. To the extent such general statement or 47 documentation includes such information, such general statement or 48 documentation shall be exempt from disclosure. To the extent any infor- 49 mation contained in such general statement or documentation is subject 50 to the attorney-client privilege or work product protection, such 51 disclosure shall not constitute a waiver of such privilege or 52 protection. 53 § 1552. Risk management. 1. (a) Beginning on January first, two thou- 54 sand twenty-seven, each deployer of a high-risk artificial intelligence 55 decision system shall use reasonable care to protect consumers from any 56 known or reasonably foreseeable risks of algorithmic discrimination. In 
 S. 1962 7 1 any enforcement action brought on or after said date by the attorney 2 general pursuant to this article, there shall be a rebuttable presump- 3 tion that a deployer of a high-risk artificial intelligence decision 4 system used reasonable care as required pursuant to this subdivision if: 5 (i) the deployer complied with the provisions of this section; and 6 (ii) an independent third party identified by the attorney general 7 pursuant to paragraph (b) of this subdivision and retained by the 8 deployer completed bias and governance audits for the high-risk artifi- 9 cial intelligence decision system. 10 (b) No later than January first, two thousand twenty-seven, and at 11 least annually thereafter, the attorney general shall: 12 (i) identify the independent third parties who, in the attorney gener- 13 al's opinion, are qualified to complete bias and governance audits for 14 the purposes of subparagraph (ii) of paragraph (a) of this subdivision; 15 and 16 (ii) make a list of such independent third parties available on the 17 attorney general's web site. 18 2. (a) Beginning on January first, two thousand twenty-seven, and 19 except as provided in subdivision seven of this section, each deployer 20 of a high-risk artificial intelligence decision system shall implement 21 and maintain a risk management policy and program to govern such 22 deployer's deployment of the high-risk artificial intelligence decision 23 system. The risk management policy and program shall specify and incor- 24 porate the principles, processes, and personnel that the deployer shall 25 use to identify, document, and mitigate any known or reasonably foresee- 26 able risks of algorithmic discrimination. The risk management policy 27 shall be the product of an iterative process, the risk management 28 program shall be an iterative process and both the risk management poli- 29 cy and program shall be planned, implemented, and regularly and system- 30 atically reviewed and updated over the lifecycle of the high-risk arti- 31 ficial intelligence decision system. Each risk management policy and 32 program implemented and maintained pursuant to this subdivision shall be 33 reasonable, considering: 34 (i) the guidance and standards set forth in the latest version of: 35 (A) the "Artificial Intelligence Risk Management Framework" published 36 by the national institute of standards and technology; 37 (B) ISO or IEC 42001 of the international organization for standardi- 38 zation; or 39 (C) a nationally or internationally recognized risk management frame- 40 work for artificial intelligence decision systems, other than the guid- 41 ance and standards specified in clauses (A) and (B) of this subpara- 42 graph, that imposes requirements that are substantially equivalent to, 43 and at least as stringent as, the requirements established pursuant to 44 this section for risk management policies and programs; 45 (ii) the size and complexity of the deployer; 46 (iii) the nature and scope of the high-risk artificial intelligence 47 decision systems deployed by the deployer, including, but not limited 48 to, the intended uses of such high-risk artificial intelligence decision 49 systems; and 50 (iv) the sensitivity and volume of data processed in connection with 51 the high-risk artificial intelligence decision systems deployed by the 52 deployer. 53 (b) A risk management policy and program implemented and maintained 54 pursuant to paragraph (a) of this subdivision may cover multiple high- 55 risk artificial intelligence decision systems deployed by the deployer. 
 S. 1962 8 1 3. (a) Except as provided in paragraphs (c) and (d) of this subdivi- 2 sion and subdivision seven of this section: 3 (i) a deployer that deploys a high-risk artificial intelligence deci- 4 sion system on or after January first, two thousand twenty-seven, or a 5 third party contracted by the deployer, shall complete an impact assess- 6 ment of the high-risk artificial intelligence decision system; and 7 (ii) beginning on January first, two thousand twenty-seven, a deploy- 8 er, or a third party contracted by the deployer, shall complete an 9 impact assessment of a deployed high-risk artificial intelligence deci- 10 sion system: 11 (A) at least annually; and 12 (B) no later than ninety days after an intentional and substantial 13 modification to such high-risk artificial intelligence decision system 14 is made available. 15 (b) (i) Each impact assessment completed pursuant to this subdivision 16 shall include, at a minimum and to the extent reasonably known by, or 17 available to, the deployer: 18 (A) a statement by the deployer disclosing the purpose, intended use 19 cases and deployment context of, and benefits afforded by, the high-risk 20 artificial intelligence decision system; 21 (B) an analysis of whether the deployment of the high-risk artificial 22 intelligence decision system poses any known or reasonably foreseeable 23 risks of algorithmic discrimination and, if so, the nature of such algo- 24 rithmic discrimination and the steps that have been taken to mitigate 25 such risks; 26 (C) A description of: 27 (I) the categories of data the high-risk artificial intelligence deci- 28 sion system processes as inputs; and 29 (II) the outputs such high-risk artificial intelligence decision 30 system produces; 31 (D) if the deployer used data to customize the high-risk artificial 32 intelligence decision system, an overview of the categories of data the 33 deployer used to customize such high-risk artificial intelligence deci- 34 sion system; 35 (E) any metrics used to evaluate the performance and known limitations 36 of the high-risk artificial intelligence decision system; 37 (F) a description of any transparency measures taken concerning the 38 high-risk artificial intelligence decision system, including, but not 39 limited to, any measures taken to disclose to a consumer that such high- 40 risk artificial intelligence decision system is in use when such high- 41 risk artificial intelligence decision system is in use; and 42 (G) a description of the post-deployment monitoring and user safe- 43 guards provided concerning such high-risk artificial intelligence deci- 44 sion system, including, but not limited to, the oversight, use, and 45 learning process established by the deployer to address issues arising 46 from deployment of such high-risk artificial intelligence decision 47 system. 48 (ii) In addition to the statement, analysis, descriptions, overview, 49 and metrics required pursuant to subparagraph (i) of this paragraph, an 50 impact assessment completed pursuant to this subdivision following an 51 intentional and substantial modification made to a high-risk artificial 52 intelligence decision system on or after January first, two thousand 53 twenty-seven, shall include a statement disclosing the extent to which 54 the high-risk artificial intelligence decision system was used in a 55 manner that was consistent with, or varied from, the developer's 56 intended uses of such high-risk artificial intelligence decision system. 
 S. 1962 9 1 (c) A single impact assessment may address a comparable set of high- 2 risk artificial intelligence decision systems deployed by a deployer. 3 (d) If a deployer, or a third party contracted by the deployer, 4 completes an impact assessment for the purpose of complying with another 5 applicable law or regulation, such impact assessment shall be deemed to 6 satisfy the requirements established in this subdivision if such impact 7 assessment is reasonably similar in scope and effect to the impact 8 assessment that would otherwise be completed pursuant to this subdivi- 9 sion. 10 (e) A deployer shall maintain the most recently completed impact 11 assessment of a high-risk artificial intelligence decision system as 12 required pursuant to this subdivision, all records concerning each such 13 impact assessment and all prior impact assessments, if any, for a period 14 of at least three years following the final deployment of the high-risk 15 artificial intelligence decision system. 16 4. Except as provided in subdivision seven of this section, a deploy- 17 er, or a third party contracted by the deployer, shall review, no later 18 than January first, two thousand twenty-seven, and at least annually 19 thereafter, the deployment of each high-risk artificial intelligence 20 decision system deployed by the deployer to ensure that such high-risk 21 artificial intelligence decision system is not causing algorithmic 22 discrimination. 23 5. (a) Beginning on January first, two thousand twenty-seven, and 24 before a deployer deploys a high-risk artificial intelligence decision 25 system to make, or be a substantial factor in making, a consequential 26 decision concerning a consumer, the deployer shall: 27 (i) notify the consumer that the deployer has deployed a high-risk 28 artificial intelligence decision system to make, or be a substantial 29 factor in making, such consequential decision; and 30 (ii) provide to the consumer: 31 (A) a statement disclosing: 32 (I) the purpose of such high-risk artificial intelligence decision 33 system; and 34 (II) the nature of such consequential decision; 35 (B) contact information for such deployer; 36 (C) a description, in plain language, of such high-risk artificial 37 intelligence decision system; and 38 (D) instructions on how to access the statement made available pursu- 39 ant to paragraph (a) of subdivision six of this section. 40 (b) Beginning on January first, two thousand twenty-seven, a deployer 41 that has deployed a high-risk artificial intelligence decision system to 42 make, or as a substantial factor in making, a consequential decision 43 concerning a consumer shall, if such consequential decision is adverse 44 to the consumer, provide to such consumer: 45 (i) a statement disclosing the principal reason or reasons for such 46 adverse consequential decision, including, but not limited to: 47 (A) the degree to which, and manner in which, the high-risk artificial 48 intelligence decision system contributed to such adverse consequential 49 decision; 50 (B) the type of data that was processed by such high-risk artificial 51 intelligence decision system in making such adverse consequential deci- 52 sion; and 53 (C) the source of such data; and 54 (ii) an opportunity to: 
 S. 1962 10 1 (A) correct any incorrect personal data that the high-risk artificial 2 intelligence decision system processed in making, or as a substantial 3 factor in making, such adverse consequential decision; and 4 (B) appeal such adverse consequential decision, which shall, if tech- 5 nically feasible, allow for human review unless providing such opportu- 6 nity is not in the best interest of such consumer, including, but not 7 limited to, in instances in which any delay might pose a risk to the 8 life or safety of such consumer. 9 (c) The deployer shall provide the notice, statements, information, 10 description, and instructions required pursuant to paragraphs (a) and 11 (b) of this subdivision: 12 (i) directly to the consumer; 13 (ii) in plain language; 14 (iii) in all languages in which such deployer, in the ordinary course 15 of such deployer's business, provides contracts, disclaimers, sale 16 announcements, and other information to consumers; and 17 (iv) in a format that is accessible to consumers with disabilities. 18 6. (a) Beginning on January first, two thousand twenty-seven, and 19 except as provided in subdivision seven of this section, each deployer 20 shall make available, in a manner that is clear and readily available on 21 such deployer's website, a statement summarizing: 22 (i) the types of high-risk artificial intelligence decision systems 23 that are currently deployed by such deployer; 24 (ii) how such deployer manages any known or reasonably foreseeable 25 risks of algorithmic discrimination that may arise from deployment of 26 each high-risk artificial intelligence decision system described in 27 subparagraph (i) of this paragraph; and 28 (iii) in detail, the nature, source and extent of the information 29 collected and used by such deployer. 30 (b) Each deployer shall periodically update the statement required 31 pursuant to paragraph (a) of this subdivision. 32 7. The provisions of subdivisions two, three, four, and six of this 33 section shall not apply to a deployer if, at the time the deployer 34 deploys a high-risk artificial intelligence decision system, and at all 35 times while the high-risk artificial intelligence decision system is 36 deployed: 37 (a) the deployer: 38 (i) has entered into a contract with the developer in which the devel- 39 oper has agreed to assume the deployer's duties pursuant to subdivisions 40 two, three, four, or six of this section; and 41 (ii) does not exclusively use such deployer's own data to train such 42 high-risk artificial intelligence decision system; 43 (b) such high-risk artificial intelligence decision system: 44 (i) is used for the intended uses that are disclosed to such deployer 45 pursuant to subparagraph (iv) of paragraph (b) of subdivision two of 46 section one thousand five hundred fifty-one of this article; and 47 (ii) continues learning based on a broad range of data sources and not 48 solely based on the deployer's own data; and 49 (c) such deployer makes available to consumers any impact assessment 50 that: 51 (i) the developer of such high-risk artificial intelligence decision 52 system has completed and provided to such deployer; and 53 (ii) includes information that is substantially similar to the infor- 54 mation included in the statement, analysis, descriptions, overview, and 55 metrics required pursuant to subparagraph (i) of paragraph (b) of subdi- 56 vision three of this section. 
 S. 1962 11 1 8. Nothing in this subdivision or subdivisions two, three, four, five, 2 or six of this section shall be construed to require a deployer to 3 disclose any information that is a trade secret or otherwise protected 4 from disclosure pursuant to state or federal law. If a deployer with- 5 holds any information from a consumer pursuant this subdivision, the 6 deployer shall send notice to such consumer disclosing: 7 (a) that the deployer is withholding such information from such 8 consumer; and 9 (b) the basis for the deployer's decision to withhold such information 10 from such consumer. 11 9. Beginning on January first, two thousand twenty-seven, the attorney 12 general may require that a deployer, or a third party contracted by the 13 deployer pursuant to subdivision three of this section, as applicable, 14 disclose to the attorney general, as part of an investigation conducted 15 by the attorney general, no later than ninety days after a request by 16 the attorney general, and in a form and manner prescribed by the attor- 17 ney general, the risk management policy implemented pursuant to subdivi- 18 sion two of this section, the impact assessment completed pursuant to 19 subdivision three of this section; or records maintained pursuant to 20 paragraph (e) of subdivision three of this section. The attorney general 21 may evaluate such risk management policy, impact assessment or records 22 to ensure compliance with the provisions of this section. In disclosing 23 such risk management policy, impact assessment or records to the attor- 24 ney general pursuant to this subdivision, the deployer or third-party 25 contractor, as applicable, may designate such risk management policy, 26 impact assessment or records as including any information that is exempt 27 from disclosure pursuant to subdivision eight of this section or article 28 six of the public officers law. To the extent such risk management poli- 29 cy, impact assessment, or records include such information, such risk 30 management policy, impact assessment, or records shall be exempt from 31 disclosure. To the extent any information contained in such risk manage- 32 ment policy, impact assessment, or record is subject to the attorney- 33 client privilege or work product protection, such disclosure shall not 34 constitute a waiver of such privilege or protection. 35 § 1553. Technical documentation. 1. Beginning on January first, two 36 thousand twenty-seven, each developer of a general-purpose artificial 37 intelligence model shall, except as provided in subdivision two of this 38 section: 39 (a) create and maintain technical documentation for the general-pur- 40 pose artificial intelligence model, which shall: 41 (i) include: 42 (A) the training and testing processes for such general-purpose arti- 43 ficial intelligence model; and 44 (B) the results of an evaluation of such general-purpose artificial 45 intelligence model performed to determine whether such general-purpose 46 artificial intelligence model is in compliance with the provisions of 47 this article; 48 (ii) include, as appropriate, considering the size and risk profile of 49 such general-purpose artificial intelligence model, at least: 50 (A) the tasks such general-purpose artificial intelligence model is 51 intended to perform; 52 (B) the type and nature of artificial intelligence decision systems in 53 which such general-purpose artificial intelligence model is intended to 54 be integrated; 55 (C) acceptable use policies for such general-purpose artificial intel- 56 ligence model; 
 S. 1962 12 1 (D) the date such general-purpose artificial intelligence model is 2 released; 3 (E) the methods by which such general-purpose artificial intelligence 4 model is distributed; and 5 (F) the modality and format of inputs and outputs for such general- 6 purpose artificial intelligence model; and 7 (iii) be reviewed and revised at least annually, or more frequently, 8 as necessary to maintain the accuracy of such technical documentation; 9 and 10 (b) create, implement, maintain and make available to persons that 11 intend to integrate such general-purpose artificial intelligence model 12 into such persons' artificial intelligence decision systems documenta- 13 tion and information that: 14 (i) enables such persons to: 15 (A) understand the capabilities and limitations of such general-pur- 16 pose artificial intelligence model; and 17 (B) comply with such persons' obligations pursuant to this article; 18 (ii) discloses, at a minimum: 19 (A) the technical means required for such general-purpose artificial 20 intelligence model to be integrated into such persons' artificial intel- 21 ligence decision systems; 22 (B) the information listed in subparagraph (ii) of paragraph (a) of 23 this subdivision; and 24 (iii) except as provided in subdivision two of this section, is 25 reviewed and revised at least annually, or more frequently, as necessary 26 to maintain the accuracy of such documentation and information. 27 2. (a) The provisions of paragraph (a) and subparagraph (iii) of para- 28 graph (b) of subdivision one of this section shall not apply to a devel- 29 oper that develops, or intentionally and substantially modifies, a 30 general-purpose artificial intelligence model on or after January first, 31 two thousand twenty-seven, if: 32 (i) (A) the developer releases such general-purpose artificial intel- 33 ligence model under a free and open-source license that allows for: 34 (I) access to, and modification, distribution, and usage of, such 35 general-purpose artificial intelligence model; and 36 (II) the parameters of such general-purpose artificial intelligence 37 model to be made publicly available pursuant to clause (B) of this 38 subparagraph; and 39 (B) unless such general-purpose artificial intelligence model is 40 deployed as a high-risk artificial intelligence decision system, the 41 parameters of such general-purpose artificial intelligence model, 42 including, but not limited to, the weights and information concerning 43 the model architecture and model usage for such general-purpose artifi- 44 cial intelligence model, are made publicly available; or 45 (ii) the general-purpose artificial intelligence model is: 46 (A) not offered for sale in the market; 47 (B) not intended to interact with consumers; and 48 (C) solely utilized: 49 (I) for an entity's internal purposes; or 50 (II) pursuant to an agreement between multiple entities for such enti- 51 ties' internal purposes. 52 (b) The provisions of this section shall not apply to a developer that 53 develops, or intentionally and substantially modifies, a general-purpose 54 artificial intelligence model on or after January first, two thousand 55 twenty-seven, if such general purpose artificial intelligence model 56 performs tasks exclusively related to an entity's internal management 
 S. 1962 13 1 affairs, including, but not limited to, ordering office supplies or 2 processing payments. 3 (c) A developer that takes any action under an exemption pursuant to 4 paragraph (a) or (b) of this subdivision shall bear the burden of demon- 5 strating that such action qualifies for such exemption. 6 (d) A developer that is exempt pursuant to subparagraph (ii) of para- 7 graph (a) of this subdivision shall establish and maintain an artificial 8 intelligence risk management framework, which shall: 9 (i) be the product of an iterative process and ongoing efforts; and 10 (ii) include, at a minimum: 11 (A) an internal governance function; 12 (B) a map function that shall establish the context to frame risks; 13 (C) a risk management function; and 14 (D) a function to measure identified risks by assessing, analyzing and 15 tracking such risks. 16 3. Nothing in subdivision one of this section shall be construed to 17 require a developer to disclose any information that is a trade secret 18 or otherwise protected from disclosure pursuant to state or federal law. 19 4. Beginning on January first, two thousand twenty-seven, the attorney 20 general may require that a developer disclose to the attorney general, 21 as part of an investigation conducted by the attorney general, no later 22 than ninety days after a request by the attorney general and in a form 23 and manner prescribed by the attorney general, any documentation main- 24 tained pursuant to this section. The attorney general may evaluate such 25 documentation to ensure compliance with the provisions of this section. 26 In disclosing any documentation to the attorney general pursuant to this 27 subdivision, the developer may designate such documentation as including 28 any information that is exempt from disclosure pursuant to subdivision 29 three of this section or article six of the public officers law. To the 30 extent such documentation includes such information, such documentation 31 shall be exempt from disclosure. To the extent any information contained 32 in such documentation is subject to the attorney-client privilege or 33 work product protection, such disclosure shall not constitute a waiver 34 of such privilege or protection. 35 § 1554. Required disclosure. 1. Beginning on January first, two thou- 36 sand twenty-seven, and except as provided in subdivision two of this 37 section, each person doing business in this state, including, but not 38 limited to, each deployer that deploys, offers, sells, leases, licenses, 39 gives, or otherwise makes available, as applicable, any artificial 40 intelligence decision system that is intended to interact with consumers 41 shall ensure that it is disclosed to each consumer who interacts with 42 such artificial intelligence decision system that such consumer is 43 interacting with an artificial intelligence decision system. 44 2. No disclosure shall be required pursuant to subdivision one of this 45 section under circumstances in which a reasonable person would deem it 46 obvious that such person is interacting with an artificial intelligence 47 decision system. 48 § 1555. Preemption. 1. Nothing in this article shall be construed to 49 restrict a developer's, deployer's, or other person's ability to: 50 (a) comply with federal, state or municipal law; 51 (b) comply with a civil, criminal or regulatory inquiry, investi- 52 gation, subpoena, or summons by a federal, state, municipal, or other 53 governmental authority; 54 (c) cooperate with a law enforcement agency concerning conduct or 55 activity that the developer, deployer, or other person reasonably and in 56 good faith believes may violate federal, state, or municipal law; 
 S. 1962 14 1 (d) investigate, establish, exercise, prepare for, or defend a legal 2 claim; 3 (e) take immediate steps to protect an interest that is essential for 4 the life or physical safety of a consumer or another individual; 5 (f) (i) by any means other than facial recognition technology, 6 prevent, detect, protect against, or respond to: 7 (A) a security incident; 8 (B) a malicious or deceptive activity; or 9 (C) identity theft, fraud, harassment or any other illegal activity; 10 (ii) investigate, report, or prosecute the persons responsible for any 11 action described in subparagraph (i) of this paragraph; or 12 (iii) preserve the integrity or security of systems; 13 (g) engage in public or peer-reviewed scientific or statistical 14 research in the public interest that: 15 (i) adheres to all other applicable ethics and privacy laws; and 16 (ii) is conducted in accordance with: 17 (A) part forty-six of title forty-five of the code of federal regu- 18 lations, as amended; or 19 (B) relevant requirements established by the federal food and drug 20 administration; 21 (h) conduct research, testing, and development activities regarding an 22 artificial intelligence decision system or model, other than testing 23 conducted pursuant to real world conditions, before such artificial 24 intelligence decision system or model is placed on the market, deployed, 25 or put into service, as applicable; 26 (i) effectuate a product recall; 27 (j) identify and repair technical errors that impair existing or 28 intended functionality; or 29 (k) assist another developer, deployer, or person with any of the 30 obligations imposed pursuant to this article. 31 2. The obligations imposed on developers, deployers, or other persons 32 pursuant to this article shall not apply where compliance by the devel- 33 oper, deployer, or other person with the provisions of this article 34 would violate an evidentiary privilege pursuant to state law. 35 3. Nothing in this article shall be construed to impose any obligation 36 on a developer, deployer, or other person that adversely affects the 37 rights or freedoms of any person, including, but not limited to, the 38 rights of any person: 39 (a) to freedom of speech or freedom of the press guaranteed in: 40 (i) the first amendment to the United States constitution; and 41 (ii) section eight of the New York state constitution; or 42 (b) pursuant to section seventy-nine-h of the civil rights law. 43 4. Nothing in this article shall be construed to apply to any develop- 44 er, deployer, or other person: 45 (a) insofar as such developer, deployer or other person develops, 46 deploys, puts into service, or intentionally and substantially modifies, 47 as applicable, a high-risk artificial intelligence decision system: 48 (i) that has been approved, authorized, certified, cleared, developed, 49 or granted by: 50 (A) a federal agency, including, but not limited to, the federal food 51 and drug administration or the federal aviation administration, acting 52 within the scope of such federal agency's authority; or 53 (B) a regulated entity subject to supervision and regulation by the 54 federal housing finance agency; or 55 (ii) in compliance with standards that are: 56 (A) established by: 
 S. 1962 15 1 (I) any federal agency, including, but not limited to, the federal 2 office of the national coordinator for health information technology; or 3 (II) a regulated entity subject to supervision and regulation by the 4 federal housing finance agency; and 5 (B) substantially equivalent to, and at least as stringent as, the 6 standards established pursuant to this article; 7 (b) conducting research to support an application: 8 (i) for approval or certification from any federal agency, including, 9 but not limited to, the federal food and drug administration, the feder- 10 al aviation administration, or the federal communications commission; or 11 (ii) that is otherwise subject to review by any federal agency; 12 (c) performing work pursuant to, or in connection with, a contract 13 with the federal department of commerce, the federal department of 14 defense, or the national aeronautics and space administration, unless 15 such developer, deployer, or other person is performing such work on a 16 high-risk artificial intelligence decision system that is used to make, 17 or as a substantial factor in making, a decision concerning employment 18 or housing; or 19 (d) that is a covered entity, as defined by the health insurance 20 portability and accountability act of 1996 and the regulations promul- 21 gated thereunder, as amended, and providing health care recommendations 22 that: 23 (i) are generated by an artificial intelligence decision system; 24 (ii) require a health care provider to take action to implement such 25 recommendations; and 26 (iii) are not considered to be high risk. 27 5. Nothing in this article shall be construed to apply to any artifi- 28 cial intelligence decision system that is acquired by or for the federal 29 government or any federal agency or department, including, but not 30 limited to, the federal department of commerce, the federal department 31 of defense, or the national aeronautics and space administration, unless 32 such artificial intelligence decision system is a high-risk artificial 33 intelligence decision system that is used to make, or as a substantial 34 factor in making, a decision concerning employment or housing. 35 6. Any insurer, as defined by section five hundred one of the insur- 36 ance law, or fraternal benefit society, as defined by section four thou- 37 sand five hundred one of the insurance law, shall be deemed to be in 38 full compliance with the provisions of this article if such insurer or 39 fraternal benefit society has implemented and maintains a written arti- 40 ficial intelligence decision systems program in accordance with all 41 requirements established by the superintendent of financial services. 42 7. (a) Any bank, out-of-state bank, New York credit union, federal 43 credit union, or out-of-state credit union, or any affiliate or subsid- 44 iary thereof, shall be deemed to be in full compliance with the 45 provisions of this article if such bank, out-of-state bank, New York 46 credit union, federal credit union, out-of-state credit union, affil- 47 iate, or subsidiary is subject to examination by any state or federal 48 prudential regulator pursuant to any published guidance or regulations 49 that apply to the use of high-risk artificial intelligence decision 50 systems, and such guidance or regulations: 51 (i) impose requirements that are substantially equivalent to, and at 52 least as stringent as, the requirements of this article; and 53 (ii) at a minimum, require such bank, out-of-state bank, New York 54 credit union, federal credit union, out-of-state credit union, affil- 55 iate, or subsidiary to: 
 S. 1962 16 1 (A) regularly audit such bank's, out-of-state bank's, New York credit 2 union's, federal credit union's, out-of-state credit union's, affil- 3 iate's, or subsidiary's use of high-risk artificial intelligence deci- 4 sion systems for compliance with state and federal anti-discrimination 5 laws and regulations applicable to such bank, out-of-state bank, New 6 York credit union, federal credit union, out-of-state credit union, 7 affiliate, or subsidiary; and 8 (B) mitigate any algorithmic discrimination caused by the use of a 9 high-risk artificial intelligence decision system, or any risk of algo- 10 rithmic discrimination that is reasonably foreseeable as a result of the 11 use of a high-risk artificial intelligence decision system. 12 (b) For the purposes of this subdivision, the following terms shall 13 have the following meanings: 14 (i) "Affiliate" shall have the same meaning as set forth in section 15 nine hundred twelve of the business corporation law. 16 (ii) "Bank" shall have the same meaning as set forth in section two of 17 the banking law. 18 (iii) "Credit union" shall have the same meaning as set forth in 19 section two of the banking law. 20 (iv) "Out-of-state bank" shall have the same meaning as set forth in 21 section two hundred twenty-two of the banking law. 22 (v) "Subsidiary" shall have the same meaning as set forth in section 23 one hundred forty-one of the banking law. 24 8. If a developer, deployer, or other person engages in any action 25 under an exemption pursuant to subdivisions one, two, three, four, five, 26 six, or seven of this section, the developer, deployer, or other person 27 bears the burden of demonstrating that such action qualifies for such 28 exemption. 29 § 1556. Enforcement. 1. The attorney general shall have exclusive 30 authority to enforce the provisions of this article. 31 2. Except as provided in subdivision six of this section, during the 32 period beginning on January first, two thousand twenty-seven, and ending 33 on January first, two thousand twenty-eight, the attorney general shall, 34 prior to initiating any action for a violation of this section, issue a 35 notice of violation to the developer, deployer, or other person if the 36 attorney general determines that it is possible to cure such violation. 37 If the developer, deployer, or other person fails to cure such violation 38 within sixty days after receipt of such notice of violation, the attor- 39 ney general may bring an action pursuant to this section. 40 3. Except as provided in subdivision six of this section, beginning on 41 January first, two thousand twenty-eight, the attorney general may, in 42 determining whether to grant a developer, deployer, or other person the 43 opportunity to cure a violation described in subdivision two of this 44 section, consider: 45 (a) the number of violations; 46 (b) the size and complexity of the developer, deployer, or other 47 person; 48 (c) the nature and extent of the developer's, deployer's, or other 49 person's business; 50 (d) the substantial likelihood of injury to the public; 51 (e) the safety of persons or property; and 52 (f) whether such violation was likely caused by human or technical 53 error. 54 4. Nothing in this article shall be construed as providing the basis 55 for a private right of action for violations of the provisions of this 56 article. 
 S. 1962 17 1 5. Except as provided in subdivisions one, two, three, four, and six 2 of this section, a violation of the requirements established in this 3 article shall constitute an unfair trade practice for purposes of 4 section three hundred forty-nine of this chapter and shall be enforced 5 solely by the attorney general; provided, however, that subdivision (h) 6 of section three hundred forty-nine of this chapter shall not apply to 7 any such violation. 8 6. (a) In any action commenced by the attorney general for any 9 violation of this article, it shall be an affirmative defense that the 10 developer, deployer, or other person: 11 (i) discovers a violation of any provision of this article through 12 red-teaming; 13 (ii) no later than sixty days after discovering such violation through 14 red-teaming: 15 (A) cures such violation; and 16 (B) provides to the attorney general, in a form and manner prescribed 17 by the attorney general, notice that such violation has been cured and 18 evidence that any harm caused by such violation has been mitigated; and 19 (iii) is otherwise in compliance with the latest version of: 20 (A) the Artificial Intelligence Risk Management Framework published by 21 the national institute of standards and technology; 22 (B) ISO/IEC 42001 of the international organization for standardi- 23 zation and the international electrotechnical commission; 24 (C) a nationally or internationally recognized risk management frame- 25 work for artificial intelligence decision systems, other than the risk 26 management frameworks described in clauses (A) and (B) of this subpara- 27 graph, that imposes requirements that are substantially equivalent to, 28 and at least as stringent as, the requirements established pursuant to 29 this article; or 30 (D) any risk management framework for artificial intelligence decision 31 systems that is substantially equivalent to, and at least as stringent 32 as, the risk management frameworks described in clauses (A), (B), and 33 (C) of this subparagraph. 34 (b) The developer, deployer, or other person bears the burden of 35 demonstrating to the attorney general that the requirements established 36 pursuant to paragraph (a) of this subdivision have been satisfied. 37 (c) Nothing in this article, including, but not limited to, the 38 enforcement authority granted to the attorney general pursuant to this 39 section, shall be construed to preempt or otherwise affect any right, 40 claim, remedy, presumption, or defense available at law or in equity. 41 Any rebuttable presumption or affirmative defense established pursuant 42 to this article shall apply only to an enforcement action brought by the 43 attorney general pursuant to this section and shall not apply to any 44 right, claim, remedy, presumption, or defense available at law or in 45 equity. 46 § 3. This act shall take effect on the two hundred seventieth day 47 after it shall have become a law.