30 bills across 10 states — frontier model disclosure requirements, safety incident reporting obligations, and whistleblower protections for AI safety researchers. California and New York have enacted laws; other states are developing proposals.
Editorially curated and verified against primary sources. Only bills with verified legislative status are included.
| State | Bill | Status | Effective | Category | Source |
|---|---|---|---|---|---|
| Frontier Model Disclosure | |||||
| California | SB 53 (Transparency in Frontier AI Act) | Enacted | 2026-01-01 | Frontier Model Disclosure | Primary Source |
| New York | RAISE Act (S 6953-B) | Enacted (Dec 2025, amended March 27, 2026) | 2026 | Frontier Model Disclosure | Primary Source |
| State | Bill | Status | Effective | Category | Summary |
|---|---|---|---|---|---|
| California | SB 53 (Transparency in Frontier AI Act) | Enacted | 2026-01-01 | Frontier Model Disclosure | Frontier model developers (>10^26 ops compute) must publish transparency reports, report critical safety incidents within 15 days, follow whistleblower procedures. Large developers (>$500M revenue) face additional safety framework requirements. Civil penalty up to $1M/violation. |
| California | AB 2653 | In Committee | Frontier Model Disclosure | Regulates frontier AI model development and deployment | |
| Connecticut | SB00005 | Introduced | Frontier Model Disclosure | Requires transparency reporting for frontier AI developers | |
| Idaho | S 1297 | Enacted | 2027-07-01 | Frontier Model Disclosure | Regulates frontier AI model development and deployment |
| Illinois | SB 3261 | In Committee | 2027-01-01 | Safety Incident Reporting | Requires safety incident disclosure and whistleblower protections |
| Illinois | SB 3444 | In Committee | Frontier Model Disclosure | Requires transparency reporting for frontier AI developers | |
| Illinois | SB3261 | Introduced | 2027-01-01 | Safety Incident Reporting | Requires safety incident disclosure and whistleblower protections |
| Illinois | SB3312 | Introduced | Safety Incident Reporting | Requires safety incident disclosure and whistleblower protections | |
| Illinois | HB4799 | Introduced | Safety Incident Reporting | Requires safety incident disclosure and whistleblower protections | |
| Illinois | HB4705 | Introduced | 2027-01-01 | Safety Incident Reporting | Requires safety incident disclosure and whistleblower protections |
| Illinois | HB3506 | Introduced | Safety Incident Reporting | Requires safety incident disclosure and whistleblower protections | |
| Illinois | SB3444 | Introduced | Frontier Model Disclosure | Requires transparency reporting for frontier AI developers | |
| Illinois | HB 4799 | Unknown | Frontier Model Disclosure | Regulates frontier AI model development and deployment | |
| Maryland | HB1399 | Introduced | Safety Incident Reporting | Requires safety incident disclosure and whistleblower protections | |
| Minnesota | HF4532 | Introduced | Frontier Model Disclosure | Requires transparency reporting for frontier AI developers | |
| Nebraska | LB525 | Passed Both Chambers | Frontier Model Disclosure | Regulates frontier AI model development and deployment | |
| New Jersey | AR158 | Passed Both Chambers | Safety Incident Reporting | Requires safety incident disclosure and whistleblower protections | |
| New Jersey | SR121 | Introduced | Safety Incident Reporting | Requires safety incident disclosure and whistleblower protections | |
| New Jersey | SR52 | Introduced | Safety Incident Reporting | Requires safety incident disclosure and whistleblower protections | |
| New York | RAISE Act (S 6953-B) | Enacted (Dec 2025, amended March 27, 2026) | 2026 | Frontier Model Disclosure | Requires frontier AI model developers operating in New York to disclose model capabilities, training data sources, and safety testing. Annual transparency reporting required. Whistleblower protections for AI safety researchers. |
| New York | S06953 | Passed Both Chambers | Compute Threshold Regulation | Regulates AI systems above compute training thresholds | |
| New York | A10583 | Introduced | 2024-09-18 | Frontier Model Disclosure | Regulates frontier AI model development and deployment |
| New York | A07278 | Introduced | 2025-06-19 | Frontier Model Disclosure | Regulates frontier AI model development and deployment |
| New York | A06453 | Introduced | Compute Threshold Regulation | Regulates AI systems above compute training thresholds | |
| United States (Federal) | HB5315 | Introduced | Frontier Model Disclosure | Regulates frontier AI model development and deployment | |
| United States (Federal) | HB3434 | Introduced | Frontier Model Disclosure | Regulates frontier AI model development and deployment | |
| United States (Federal) | HB3460 | Introduced | Safety Incident Reporting | Requires safety incident disclosure and whistleblower protections | |
| United States (Federal) | SB1775 | Introduced | Frontier Model Disclosure | Regulates frontier AI model development and deployment | |
| United States (Federal) | SB1792 | Introduced | Safety Incident Reporting | Requires safety incident disclosure and whistleblower protections | |
| United States (Federal) | HB8094 | Introduced | Compute Threshold Regulation | Regulates AI systems above compute training thresholds |
Get alerted when new frontier AI model bills are introduced or existing ones change status.
Subscribe to Bill AlertsFree alerts · Unsubscribe anytime
Frontier AI models are the largest, most capable AI systems at the leading edge of capabilities — such as GPT-4, Claude, Gemini, and similar systems. California SB 53 defines covered frontier models as those trained with more than 1026 floating-point operations (compute). These systems are regulated separately because their capabilities and potential risks are considered qualitatively different from other AI applications.
California SB 53 (effective January 1, 2026) requires frontier model developers to: publish transparency reports covering model capabilities, limitations, and safety testing; report critical safety incidents to California authorities within 15 days; implement whistleblower protections for AI safety researchers; and for large frontier developers (over $500M revenue), adopt and follow a documented safety framework. Civil penalties up to $1 million per violation.
The New York RAISE Act (S 6953-B), enacted December 2025 and amended March 27, 2026, requires frontier AI model developers operating in New York to disclose their models’ capabilities, training data sources, and safety testing methodologies. It also requires annual transparency reporting and establishes baseline whistleblower protections for AI safety researchers.