Artificial Intelligence Policy Tracker
Search
Menu

Frontier AI laws: California SB 53, New York's RAISE Act and EU systemic-risk duties

SB 53, the RAISE Act and the EU systemic-risk duties ask frontier developers for the same core set: a published safety framework, incident reporting, transparency and protection for whistleblowers.

  1. Who is a frontier developer

    Each law sets its own thresholds, by training compute and, for some duties, revenue. In the EU, a general-purpose model is presumed to have systemic risk when its cumulative training compute exceeds 10^25 floating-point operations, and the provider must notify the Commission within two weeks of meeting the threshold.

  2. A published safety framework

    California SB 53 requires large frontier developers to publish a frontier AI framework describing how they assess and mitigate catastrophic risk. New York's RAISE Act requires a safety and security protocol. The EU asks systemic-risk providers to assess and mitigate systemic risks, including through adversarial testing.

  3. Transparency

    Under SB 53, frontier developers publish a transparency report before deploying a new frontier model, and large developers send periodic summaries of catastrophic-risk assessments to the state.

  4. Incident reporting

    SB 53 requires critical safety incidents to be reported to the California Office of Emergency Services; RAISE requires reporting to New York within a set period; the EU requires systemic-risk providers to track and report serious incidents to the AI Office.

  5. Whistleblowers and security

    SB 53 protects employees who report catastrophic-risk concerns. The EU requires systemic-risk providers to secure the model and its infrastructure.

  6. Dates and confidence

    SB 53 has applied since 1 January 2026 (record verified). The RAISE Act applies from 1 January 2027; its record is verified but at low confidence, so check the text. The EU general-purpose AI duties have applied since 2 August 2025. California's September 2026 executive order on a shutdown capability creates no obligations yet.

Obligations referenced in this guide

Frequently asked questions

Does California SB 53 apply to my company?

Only if you develop frontier AI models above its compute threshold; several duties apply only to large frontier developers above a revenue threshold. Check the definitions in the statute.

When does New York's RAISE Act take effect?

On 1 January 2027, according to the record, which is verified at low confidence.

Where are frontier AI incidents reported?

In California, to the Office of Emergency Services under SB 53; in New York, to the state under the RAISE Act; in the EU, systemic-risk model providers report serious incidents to the AI Office.

What is the EU threshold for systemic-risk models?

A general-purpose model is presumed to have systemic risk when its cumulative training compute exceeds 10^25 floating-point operations, or when the Commission designates it.

Disclaimer: informational only, not legal advice. Verify every claim against the linked official sources and consult a qualified lawyer before acting.