New York General Business Law, Article 44-B — the Responsible AI Safety and Education (RAISE) Act — as enacted by the S.8828 chapter amendment, signed March 27, 2026. The chapter amendment repealed the original Article 44-B and replaced it with §§ 1420–1429. The regime takes effect January 1, 2027.
The full text below is drawn from the enacted bill (S.8828). This is an independent reference rendering for readability; the controlling text is the signed bill itself.
§ 1420 — Definitions
Section 1420 fixes the vocabulary the rest of Article 44-B runs on. A frontier model is a foundation model trained using more than 10^26 integer or floating-point operations (counting the original training run plus any subsequent fine-tuning or other modification). A foundation model is an AI model trained on a broad data set, designed for generality of output, and adaptable to a wide range of distinct tasks. A frontier developer is a person who has trained — or initiated the training of — a frontier model; a large frontier developer is a frontier developer that, with its affiliates, had annual gross revenues over $500 million in the preceding calendar year. Catastrophic risk and critical safety incident are defined here too (see § 1421/§ 1422). 'Office' means the new office within the Department of Financial Services tasked with implementing the article.
Frontier model: Foundation model > 10²⁶ operations · Large frontier developer: > $500M annual gross revenue (w/ affiliates) · Catastrophic risk: > 50 deaths/serious injuries OR > $1B in damage
§ 1421 — Transparency requirements
Section 1421 is the heart of the article. A large frontier developer must write, implement, comply with, and conspicuously publish a frontier AI framework describing how it: incorporates national/international standards; defines thresholds to identify whether a model poses catastrophic risk; applies and reviews mitigations; uses third parties to assess risk; updates the framework; secures unreleased model weights; institutes internal governance; identifies and responds to critical safety incidents; and assesses catastrophic risk from internal use. The framework is reviewed at least annually, and a material modification must be published with a justification within thirty days. Separately, before deploying a new or substantially modified frontier model, a frontier developer must publish a transparency report (developer contact, release date, supported languages and output modalities, intended uses, use restrictions, and summaries of the catastrophic-risk assessments). Publishing this in a system card or model card is deemed compliant. Redactions are allowed to protect trade secrets, cybersecurity, public safety, or national security, with the unredacted version retained for five years.
Framework: Publish; review ≥ annually; material change in 30 days · Transparency report: Before each deploy (system/model card OK) · Redactions: Allowed; retain unredacted 5 years
§ 1422 — Reporting
Section 1422 sets the incident-reporting clocks. A frontier developer must report any critical safety incident to the Office within seventy-two hours of determining one occurred — or of learning facts sufficient to establish a reasonable belief that one occurred. If the developer discovers that an incident poses an imminent risk of death or serious physical injury, it must disclose to an appropriate law-enforcement or public-safety agency within twenty-four hours (and still report to the Office within 72 hours). Separately, a large frontier developer must transmit to the Office a summary of any assessment of catastrophic risk from internal use of its frontier models every three months. The Office must allow confidential submission and limit access. The Office may share reports (including with the Attorney General), may designate equivalent or stricter federal reporting regimes as compliance methods, and — beginning January 1, 2028 — must publish an annual anonymized/aggregated report to the Governor and Legislature.
Critical safety incident: 72 hours to the Office · Imminent risk: 24 hours to law enforcement · Internal-use assessment: Every 3 months (large devs)
§ 1423 — Loss of equity
Section 1423 is a short interpretive rule. It clarifies that 'the loss of value of equity shall not count as damage to or loss of property for the purposes of this article.' In other words, when the catastrophic-risk threshold turns on more than one billion dollars in property damage or loss, a drop in share or equity value is not the kind of loss that counts — the harm must be to property itself.
Rule: Equity-value loss ≠ property damage
§ 1424 — Duties and obligations
Section 1424 is a savings clause. The duties and obligations imposed by Article 44-B are cumulative with any other duties or obligations imposed under other law; the article does not relieve any party from those other duties, and it does not limit any rights or remedies otherwise available under existing law. The RAISE Act sits on top of — not in place of — existing legal obligations.
Effect: Cumulative; no carve-out from other law
§ 1425 — Scope
Section 1425 draws the jurisdictional boundary: Article 44-B applies only to frontier models that are developed, deployed, or operating in whole or in part in New York State. The nexus to New York is what brings a model — and its developer — within reach of the regime.
Reach: Models developed/deployed/operating in NY
§ 1426 — Exceptions
Section 1426 carves out two categories. The article does not apply to accredited colleges and universities in New York State to the extent they are engaging in academic research regarding AI models, nor to the Empire AI consortium or institute (as defined in section 361 of the economic development law). Academic and public-research work is excluded from the developer obligations.
Excluded: NY academic research · Empire AI
§ 1427 — Violations
Section 1427 vests enforcement in the New York Attorney General, who may bring a civil action for penalties not exceeding one million dollars for a first violation and not exceeding three million dollars per subsequent violation, scaled to severity. Violations include a large frontier developer's failure to publish a compliant framework or transparency report, a materially false statement in violation of § 1421(4), failure to report an incident under § 1422, or failure to comply with its own frontier AI framework. The article creates no private right of action, and nothing in it prevents a developer from asserting that another person, entity, or factor was responsible for a harm.
Enforcer: NY Attorney General (no private suit) · Penalties: ≤ $1M first · ≤ $3M subsequent
§ 1428 — Large frontier developer disclosure
Section 1428 is the registration regime. No large frontier developer may develop, deploy, or operate a frontier model in New York without a current disclosure statement on file with the Office and paying its required share. The statement is renewed every two years (or sooner on a transfer or material change) and must identify the developer and all its business names; the address of its principal place of business and each NY office; beneficial owners of a 5%-or-greater interest (or 50%+ if publicly traded), including owners over the prior five years for privately held developers; and points of contact. Large frontier developers are assessed pro-rata to fund the Office. Operating without a filing draws a $1,000-per-day civil penalty (plus the assessments owed). The Office maintains and publishes the list of large frontier developers who have filed — the public registry this project models. No such official portal exists yet; Frontier is an independent reference implementation, not the Office or any state system.
File: Disclosure statement; renew every 2 years · Discloses: Names · NY addresses · 5%+ owners · contacts · Office publishes: The public list of filers
§ 1429 — Rulemaking authority
Section 1429 grants the Office authority to adopt rules and regulations to implement Article 44-B as needed. Where the Office determines it will facilitate safety and transparency consistent with the article's purpose, it may consider additional reporting or publication requirements — including, but not limited to, post-critical-safety-incident information, sharing plans and protocols, and the transmission of frontier AI frameworks to the Office. Because much operational detail is delegated to this rulemaking, Frontier keeps its disclosure model loosely coupled / config-driven rather than hardcoding a filing format that DFS has not yet defined.
Authority: Office adopts implementing rules · May add: Reporting/publication reqs; framework transmittal