Quick Answer
New York’s RAISE Portal is now open for large frontier AI developers ahead of the state’s January 1, 2027 safety-law deadline. The portal centralizes registration and required incident reporting for covered companies. Most consumers do not need to take action, but New Yorkers should expect more public safety and transparency information from qualifying AI developers after the law takes effect.
Key Takeaways
- New York opened the RAISE Portal on November 10, 2026.
- Large frontier AI developers must register by January 1, 2027.
- The RAISE Act takes effect on January 1, 2027.
- Covered developers must report critical safety incidents within 72 hours.
- Large developers must publish safety frameworks, while all covered developers must publish transparency reports.
What is New York’s RAISE Portal?
New York’s RAISE Portal is the state’s new reporting and registration system for developers of certain advanced general-purpose AI models. The Office of Digital Innovation, Governance, Integrity, and Trust, known as DIGIT, opened the portal on November 10, 2026, as part of the implementation of the Responsible AI Safety and Education Act. New York’s announcement says the portal is intended for large frontier AI developers operating in the state.
The RAISE Portal matters because it gives New York one defined channel for registrations and mandatory safety reporting rather than leaving companies to use separate agency processes. The New York Department of Financial Services states that reporting required under the RAISE Act, including critical safety incidents, must go through the RAISE Portal instead of its standard consumer complaint form. The DFS notice clarifies the reporting path for affected companies.
The RAISE Portal does not create a new consumer account, and ordinary users of chatbots, search tools, or AI-powered apps do not need to register. The practical change is directed at qualifying developers, although public disclosures required by the law could give consumers more information when deciding whether to trust a major AI service.
When do the New York RAISE Act requirements begin?
The New York RAISE Act takes effect on January 1, 2027, and large frontier AI developers must register through the RAISE Portal by that date. The deadline is important because the portal opened less than 2 months before the law’s effective date, leaving covered companies a limited implementation window. The enacted bill text sets January 1, 2027 as the effective date.
New York’s implementation schedule also requires qualifying large frontier developers to publish safety frameworks on their websites beginning in January. All covered frontier developers, including those that do not meet the law’s large-developer revenue threshold, must publish model transparency reports. Those disclosures could make it easier for researchers, customers, and regulators to understand a model’s stated safeguards and development practices.
The RAISE Act requirements are not a guarantee that an AI product is safe or accurate. Public safety frameworks describe a developer’s processes, but they do not replace independent testing, security review, or careful human oversight. Consumers should continue to avoid entering sensitive financial, health, employment, or identity information into AI tools unless the service’s data practices are clear.
Which AI companies must use the RAISE Portal?
The New York RAISE Act applies to developers of frontier models, which the law defines as general-purpose foundation models trained using more than 1026 integer or floating-point operations. A foundation model is an AI system trained for broad uses that can later support tasks such as writing, coding, analysis, or image generation. The computing threshold is designed to focus the law on the most resource-intensive model development rather than smaller, specialized software tools.
The law adds further obligations for a “large frontier developer,” defined as a developer whose annual gross revenue, including the revenue of affiliates, exceeded $500 million in the preceding calendar year. Large frontier developers must register through the portal and publish safety frameworks. The revenue test means the law distinguishes between advanced-model developers based on organizational scale as well as model capability.
| Developer category | Threshold in the RAISE Act | Key requirement |
|---|---|---|
| Frontier developer | Develops a general-purpose foundation model trained with more than 1026 operations | Publish a model transparency report and report critical safety incidents |
| Large frontier developer | Frontier developer with more than $500 million in annual gross revenue, including affiliates | Register through the RAISE Portal and publish a safety framework |
| Consumer AI user | Uses an AI service but does not develop a covered model | No registration requirement under the RAISE Act |
The New York RAISE Act does not automatically cover every app that includes an AI feature. A photo editor, note-taking app, or customer-service tool may use AI without developing a covered frontier model. Consumers evaluating AI features in creative tools should still consider what data the app collects, especially as services expand AI-powered editing and automation.
What must covered AI developers report under the RAISE Act?
Covered frontier AI developers must report critical safety incidents within 72 hours under New York’s implementation plan. New York announced the reporting requirement when it opened the RAISE Portal on November 10, 2026. The 72-hour timeline matters because state officials can receive notice relatively soon after a developer identifies a serious incident involving a covered model.
The available state announcement confirms the deadline but does not establish that every reported incident will become public immediately. Reporting to a regulator and public disclosure are separate processes, particularly when an incident involves security, misuse, or an ongoing investigation. Consumers should not assume that the absence of a public notice means a company has never reported an incident.
Critical-incident reporting is one part of a broader AI safety process. A company still needs technical testing, internal escalation procedures, and clear limits on how a model is deployed. The same caution applies to AI services that connect to personal accounts or payment tools, where convenience can increase the amount of sensitive data handled by an automated system.
What safety and transparency information will AI developers publish?
Large frontier AI developers must publish safety frameworks on their websites beginning in January 2027, while all covered frontier developers must publish model transparency reports. New York’s November 10 announcement presents those documents as separate obligations. Safety frameworks generally describe the processes a company says it uses to identify and manage serious risks, while transparency reports provide information about covered models.
Public documentation can help outside researchers and enterprise customers compare stated safeguards across developers. The disclosure requirement also gives New Yorkers a clearer place to look before relying on an advanced AI system for work, education, creative projects, or other consequential tasks. Readers who use AI tools for projects should still verify whether a service’s outputs are accurate, licensed for the intended use, and appropriate for the information being shared.
The disclosures have an important limit: a published framework is not independent certification. A company can explain its policies without proving that every safeguard works in every circumstance. The most sensible approach is to treat safety and transparency reports as useful evidence, then consider the service’s privacy terms, security record, and the sensitivity of the task.
How does New York plan to evaluate advanced AI systems?
New York launched the New York Center for AI Responsibility and Research at SUNY Binghamton alongside the RAISE Portal. The state says the center will bring together more than 80 researchers to test and evaluate advanced AI systems and safety frameworks. The research effort matters because AI governance rules require technical expertise to assess whether developer claims match real-world model behavior.
The center’s work does not mean New York will evaluate every AI product that reaches consumers. The state announcement describes a research and evaluation effort focused on advanced systems and safety frameworks, not a consumer approval label for all apps. New Yorkers should therefore continue to assess individual services based on their own privacy settings, account protections, and intended use.
AI safety oversight is becoming a broader policy issue as models gain access to more personal data and more automated actions. Users should apply the same caution they would use with other connected services: enable strong account security, review app permissions, and avoid providing information that could cause harm if retained, exposed, or misinterpreted.
What should New York AI users and businesses do now?
New York AI users do not need to register through the RAISE Portal, but they should pay closer attention to safety frameworks and transparency reports when those documents begin appearing in January 2027. The disclosures can provide context about a covered developer’s stated safeguards, but they do not remove the need to use AI tools carefully. A chatbot can produce a confident answer that is incomplete, inaccurate, or unsuitable for a high-stakes decision.
New York businesses that buy or deploy AI services should identify whether a vendor develops a covered frontier model, uses one from another developer, or only provides an application built on top of a model. That distinction affects which organization has direct obligations under the RAISE Act. Businesses should also document what data employees can enter into AI systems and who reviews outputs before they are used in customer, employment, financial, legal, or health-related decisions.
- Review the privacy and data-use settings for AI tools your household or business uses.
- Check whether a developer publishes a safety framework or model transparency report after January 1, 2027.
- Keep passwords, government identification, financial account data, and confidential work documents out of general-purpose AI prompts unless the service is approved for that use.
- Verify important AI-generated answers against primary documents or qualified professionals before acting on them.
The practical response is to treat the New York RAISE Act as an added accountability measure rather than a substitute for user judgment. Security basics remain important because an AI service can still expose sensitive information if an account is compromised or a vendor suffers a breach. Understanding basic cybersecurity risks helps users make safer choices across AI platforms and other online services.
FAQ
What is the New York RAISE Act?
The New York RAISE Act is the Responsible AI Safety and Education Act, a state law that sets reporting, transparency, and safety-framework obligations for certain frontier AI developers. The law takes effect on January 1, 2027.
When must companies register through the RAISE Portal?
Large frontier AI developers operating in New York must register through the RAISE Portal by January 1, 2027. The registration obligation applies to developers that meet the law’s frontier-model definition and the more than $500 million annual revenue threshold.
Do regular AI users need to use the RAISE Portal?
No, regular AI users do not need to use the RAISE Portal. The portal is for developer registration and legally required reporting, while consumers can use future public disclosures to evaluate qualifying AI services.
What is a frontier model under the New York RAISE Act?
A frontier model under the New York RAISE Act is a general-purpose foundation model trained using more than 1026 integer or floating-point operations. The threshold focuses the law on highly compute-intensive advanced models rather than every AI-enabled app.
How quickly must AI safety incidents be reported?
Covered frontier AI developers must report critical safety incidents within 72 hours under New York’s announced implementation requirements. Required reports must go through the RAISE Portal rather than the Department of Financial Services consumer complaint form.