On July 6, 2026, SB 315, Illinois’s Artificial Intelligence Safety Measures Act (the Safety Measures Act or the Act) was signed into law, establishing what Illinois describes as the nation’s strongest framework for artificial intelligence (AI) safety, transparency and accountability.1

The Act incorporates concepts appearing in recently enacted state AI laws2, but also adds a new requirement of mandatory independent third-party audits of covered AI systems. It imposes significant obligations on “large frontier developers,” including requirements to publish a frontier AI framework, transparency report and annual disclosures, conduct independent third-party audits and report critical safety incidents.  The Act takes effect on January 1, 2027 (though some requirements will not go into effect until January 1, 2028), and any operational touchpoint within Illinois may bring a developer within the Act’s scope, meaning covered companies must prepare now to meet its new obligations.

This article attempts to help companies do just that by understanding the Act’s provisions and preparing to meet the Act's obligations. It begins with a summary of key provisions and potential steps to consider, follows with a detailed discussion of each, and concludes with practical takeaways and a look at the broader regulatory landscape.

Summary of key points to know about the Safety Measures Act

  • The Act applies to “large frontier developers” of “frontier models” developed, deployed or operating in whole or in part in Illinois:

    • A “large frontier developer” is defined as a frontier developer that, together with its affiliates, had annual gross revenues exceeding US$500 million in the preceding calendar year.3
    • A “frontier model” is defined as a foundation model trained using computing power greater than 10^26 integer or floating-point operations, including the original training run and any subsequent fine-tuning, reinforcement learning or other material modifications.4
  • Beginning January 1, 2027, large frontier developers must file a disclosure statement before developing, deploying or operating a frontier model in Illinois.5

  • Beginning January 1, 2028 (or 90 days after a developer first qualifies, whichever is later), large frontier developers must annually retain a third party to perform an independent compliance audit.6

  • Beginning January 1, 2028, large frontier developers must write, implement, comply with and conspicuously publish on their website a frontier AI framework describing how the developer approaches:

    • Incorporating national, international and industry-consensus best practices
    • Defining and assessing thresholds for identifying capabilities that could pose a catastrophic risk, and applying mitigations based on those assessments
    • Using third parties to assess the potential for catastrophic risks and the effectiveness of mitigations
    • Cybersecurity practices to secure unreleased model weights from unauthorized modification or transfer
    • Identifying and responding to critical safety incidents, instituting internal governance practices, and managing catastrophic risk from internal use of frontier models7
  • Large frontier developers must also transmit to applicable agencies a summary of catastrophic risk assessments every three months, unless the developer communicates an alternative reasonable schedule to certain state agencies.8

  • Before or concurrently with deploying a new frontier model or a substantially modified version, a frontier developer must publish a transparency report on its website.9

  • A frontier developer must report any critical safety incident to certain state agencies within 72 hours of establishing a reasonable belief that such an incident has occurred.10

  • If a critical safety incident poses an imminent risk of death or serious physical injury, the developer must disclose it within 24 hours to an appropriate authority.11

  • The Act creates confidential reporting channels and whistleblower protections for employees raising AI safety concerns.12

  • The Act does not establish a private right of action.13

  • A large frontier developer that fails to comply with the Act’s requirements is subject to a civil penalty that does not exceed US$1 million for the first violation and US$3 million per violation for subsequent violations.14

    • Additionally, if any person develops, deploys or operates a large frontier model in Illinois without a current disclosure, submits false information or fails to timely pay a required assessment, certain state agencies may levy a civil penalty of US$1,000 per day of noncompliance, plus an amount equal to the assessments owed.15
  • The Act includes an interoperability provision allowing state agencies to designate substantially equivalent or stricter federal standards, compliance with which satisfies the Act’s requirements.16

The Act’s scope and effective date

Under the Act, beginning January 1, 2027, no large frontier developer may develop, deploy or operate a frontier model in whole or in part in Illinois without a current disclosure statement filed with certain state agencies.17 Any operational touchpoint within Illinois may bring a large frontier developer within the Act’s scope.

The requirement to publish a frontier AI framework and begin annual third-party audits does not take effect until January 1, 2028, or 90 days after a developer first qualifies as a large frontier developer, whichever is later.18 Despite this staggered timeline, companies should begin developing their framework and planning for the audit requirement immediately, as these processes take time.

The Act also includes an interoperability provision wherein certain Illinois state agencies may designate federal laws, regulations or guidance imposing substantially equivalent or stricter standards, allowing compliant frontier developers to be deemed in compliance with the Act.19 Given the current absence of binding federal AI safety legislation, companies should not rely on this safe harbor without confirming that the those agencies have designated a qualifying federal framework.

The Act also preempts local regulation; the Act declares that the regulation of artificial intelligence frontier models is an exclusive power and function of the State, denying and limiting home rule powers under the Illinois Constitution.

The Act focuses on large frontier developers and the risk of “catastrophic harm”

The Act also imposes obligations on “frontier developers,” which are persons who train, or initiate the training of, a frontier model using computing power meeting the Act’s technical specifications.20 A “large frontier developer” is a frontier developer that, together with its affiliates, had annual gross revenues exceeding US$500 million in the preceding calendar year.21

A “frontier model” is a foundation model trained using computing power greater than 10^26 integer or floating-point operations, including the original training run and any subsequent fine-tuning, reinforcement learning or other material modifications.22  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 distinctive tasks.23

The Act aims to prevent “catastrophic risk,” defined as a foreseeable and material risk that a frontier developer’s development, storage, use, or deployment of a frontier model will materially contribute to the death of, or serious injury to, more than 50 people or more than US$1 billion in property damage or loss arising from a single incident involving a frontier model that:

  1. Provides expert-level assistance in the creation or release of a chemical, biological, radiological, or nuclear weapon
  2. Engages in conduct with no meaningful human oversight, intervention, or supervision that is either a cyberattack or, if committed by a human, would constitute murder, assault, extortion, or theft
  3. Evades the control of its frontier developer or user.24 Loss of value of equity does not count as damage to or loss of property.25

“Catastrophic risk” excludes risks from information that a frontier model outputs if that information is otherwise publicly accessible in substantially similar form from a source other than a foundation model, lawful federal government activity or harm caused by a frontier model in combination with other software if the frontier model did not materially contribute to the harm.26

The Act separately defines “critical safety incidents,” which trigger mandatory 72-hour (or, where applicable, 24-hour) reporting obligations. These include:

  1. Unauthorized access to, modification of or exfiltration of frontier model weights resulting in death or bodily injury
  2. Harm from the materialization of a catastrophic risk
  3. Loss of control of a frontier model causing death or bodily injury
  4. A frontier model using deceptive techniques to subvert its developer’s controls or monitoring outside a designated evaluation, in a manner demonstrating materially increased catastrophic risk.27

The Act's obligations for large frontier developers: Frontier AI framework, transparency reports, independent third-party audits and critical safety incident reporting

Disclosure statement

Beginning January 1, 2027, large frontier developers must file a disclosure statement with certain state agencies and pay the required fee before developing, deploying or operating a frontier model in Illinois.28

The statement must be renewed annually as well as whenever ownership of the frontier model is transferred or whenever there is a material change to the information reported.29

The statement must identify: the large frontier developer and all names under which it conducts business; the address of its principal place of business and each Illinois office; if the developer or its parent is privately or closely held, all persons or entities that beneficially own a 5 percent or greater interest at the time of filing and all persons who formerly beneficially owned a 5 percent or greater interest in the preceding five years; if the developer or its parent is publicly traded, all persons or entities beneficially owning a 50 percent or greater interest; and the name and contact information of a primary, secondary and tertiary contact.30

The frontier AI framework

Starting January 1, 2028, a large frontier developer must write, implement, comply with and conspicuously publish on its website a frontier AI framework (meaning a documented technical and organizational protocol to manage, assess and mitigate catastrophic risks31) applicable to its frontier models.32 The framework must describe how the developer approaches:

  1. Incorporating national standards, international standards and industry-consensus best practices
  2. Defining and assessing thresholds used to identify whether a frontier model has capabilities posing a catastrophic risk, which may include multiple-tiered thresholds
  3. Applying mitigations to address potential catastrophic risks based on those assessments
  4. Reviewing assessments and mitigation adequacy as part of the decision to deploy a frontier model or use it extensively internally
  5. Using third parties to assess the potential for catastrophic risks and the effectiveness of mitigations
  6. Revisiting and updating the frontier AI framework, including criteria triggering updates and how the developer determines when frontier models are substantially modified enough to require new disclosures
  7. Cybersecurity practices to secure unreleased model weights (numerical parameters that are adjusted through training and that help determine how inputs are transformed into outputs) from unauthorized modification or transfer
  8. Identifying and responding to critical safety incidents
  9. Instituting internal governance practices to confirm implementation of these processes
  10. Assessing and managing catastrophic risk resulting from the internal use of frontier models, including risks resulting from a frontier model circumventing oversight mechanisms33

A large frontier developer must review and, as appropriate, update its frontier AI framework at least annually, and material modifications must be conspicuously published with a justification for the modifications within 30 days.34

A large frontier developer must also transmit a summary of any catastrophic risk assessment resulting from internal use of its frontier models every three months, or on another reasonable schedule specified by the developer and communicated in writing to certain state agencies.35

The Act prohibits any frontier developer from making materially false or misleading statements about catastrophic risk from its frontier models or its management thereof, and prohibits large frontier developers from making such statements about their implementation of or compliance with their frontier AI framework.36 This prohibition does not apply to good-faith statements and statements that were reasonable under the circumstances.37

Transparency reports

Before or concurrently with deploying a new frontier model or a substantially modified version, a frontier developer must conspicuously publish on its website a transparency report containing: the developer's website, a mechanism for natural persons to communicate with the developer, the model’s release date, supported languages, output modalities, intended uses and any generally applicable restrictions or conditions on use.38

Large frontier developers must additionally include summaries of: catastrophic risk assessments conducted pursuant to the frontier AI framework; the results of those assessments; the extent of third-party evaluator involvement; and other steps taken to fulfill the frontier AI framework’s requirements for that model.39

Independent third-party audits

Perhaps the most novel feature of the Act compared to prior state AI laws is its mandatory independent third-party audit. Beginning January 1, 2028 (or 90 days after a developer first qualifies, whichever is later), a large frontier developer must annually retain a third party to perform an independent compliance audit and report.40 Neither party may have a financial interest in the other, and while the developer may compensate the auditor, it may not condition payment on the audit results.41

The developer must retain an unredacted copy of the audit report for as long as a frontier model is deployed plus five years, and, within 30 days of receiving the report, it must conspicuously publish a summary and a redacted copy on its website and transmit the redacted report to certain Illinois state agencies.42

When publishing documents under the Act, a frontier developer may redact information necessary to protect trade secrets, cybersecurity, public safety or US national security, or to comply with applicable law, provided it describes the character and justification of the redactions and retains the unredacted information for five years.43

Critical safety incident reporting

A frontier developer must report any critical safety incident pertaining to its frontier models to certain state agencies within 72 hours of establishing a reasonable belief that such an incident has occurred.44 The disclosure must include:

  1. The date of the incident
  2. The reasons it qualifies as a critical safety incident
  3. A short and plain statement describing the incident.45

If a critical safety incident poses an imminent risk of death or serious physical injury, the developer must disclose it within 24 hours to an appropriate authority, including any law enforcement or public safety agency with jurisdiction.46

Critical safety incident reports, summaries of catastrophic risk assessments from internal use, unredacted third-party audit reports and covered employee disclosures are all exempt from disclosure under the Illinois Freedom of Information Act.47

Whistleblower protections

The Act provides whistleblower protections for “covered employees,” which are employees responsible for assessing, managing or addressing the risk of critical safety incidents.48

A frontier developer may not adopt or enforce any rule, policy or contract that prevents or retaliates against a covered employee for disclosing information to certain state agencies, a federal authority or a person with authority to investigate the reported issue, where the employee reasonably believes the information reveals that the developer’s activities pose a specific and substantial danger to public health or safety resulting from a catastrophic risk, or that the developer has violated the Safety Measures Act.49

A large frontier developer must also provide a reasonable internal process through which a covered employee can anonymously disclose information if the employee believes in good faith that the developer’s activities present a specific and substantial danger resulting from a catastrophic risk or that the developer has violated the Act, including updates to the disclosing employee regarding the investigation’s status.50

A frontier developer must provide clear notice to all covered employees of their rights under the Act, either by posting such notice at all times in any workplace, or by providing annual written notice that is received and acknowledged by all covered employees.51

Enforcement of the Act: Penalties and remedies

A large frontier developer that fails to publish or transmit a required document, makes a materially false or misleading statement, fails to have a third-party audit performed, fails to report a critical safety incident or fails to comply with its own frontier AI framework is subject to a civil penalty up to US$1 million for the first violation and up to US$3 million per subsequent violation.52

Additionally, any person that develops, deploys or operates a large frontier model in Illinois without a current disclosure statement, submits false information or fails to timely pay required assessments may, after notice and hearing, be levied a civil penalty of US$1,000 per day of noncompliance.53

Civil penalties are recovered exclusively through civil actions brought by the government.54 There is no private right of action.55

Conclusion: Practical takeaways for companies

Illinois’s Safety Measures Act imposes significant obligations on the largest AI developers, including framework publication, transparency reporting, mandatory independent audits, incident reporting and whistleblower protections—all backed by civil penalties. Illinois and the other large states that have enacted similar legislation cover, combined, an estimated 40 percent of the US AI market. Given that, the Act and other similar state AI laws are likely to function as a de facto national standard until and unless federal AI legislation preempts them.56

Companies should consider taking at least the following steps:

  • Determine whether your company qualifies as a “large frontier developer” that will be subject to the Act.

  • If your company is covered by the Act:

    • File the disclosure statement by January 1, 2027.
    • Draft and publish a frontier AI framework addressing the required areas.
    • Establish an annual review cycle to review and update the frontier AI framework, publish material modifications with justification within 30 days and confirm all public-facing disclosures are accurate.
    • Prepare and publish transparency reports, including catastrophic risk assessment summaries and third-party evaluator involvement.
    • Build an incident response program to identify and escalate potential critical safety incidents so you can meet the 72-hour and 24-hour reporting deadlines.
    • Engage a qualified third-party auditor.
    • Implement whistleblower protections by deploying compliant workplace notices, establishing anonymous internal reporting for covered employees and confirming that no contract or policy improperly restricts covered employees’ reporting rights.

Our team will continue monitoring developments related to the Safety Measures Act, including any state agency guidance on disclosure procedures, designation of qualifying federal frameworks under the interoperability provision and any legislative or regulatory amendments. For advice or representation related to the Act, please contact our team.

Footnotes

  1. 1

    Gov. Pritzker Signs Nation-Leading Artificial Intelligence Safety Law, July 6, 2026.

  2. 2

    For a breakdown of similar state laws, see "The New York Responsible AI Safety and Education (RAISE) Act: What you need to know".

  3. 3

    Artificial Intelligence Safety Measures Act § 5.

  4. 4

    Id.

  5. 5

    Artificial Intelligence Safety Measures Act § 18(a).

  6. 6

    Artificial Intelligence Safety Measures Act § 10(d).

  7. 7

    Artificial Intelligence Safety Measures Act § 10(a).

  8. 8

    Artificial Intelligence Safety Measures Act § 10(e).

  9. 9

    Artificial Intelligence Safety Measures Act § 10(c)(1).

  10. 10

    Artificial Intelligence Safety Measures Act § 15(c).

  11. 11

    Artificial Intelligence Safety Measures Act § 15(c).

  12. 12

    Artificial Intelligence Safety Measures Act § 20.

  13. 13

    Artificial Intelligence Safety Measures Act § 25(d).

  14. 14

    Artificial Intelligence Safety Measures Act § 25(a).

  15. 15

    Artificial Intelligence Safety Measures Act § 18(e).

  16. 16

    Artificial Intelligence Safety Measures Act § 17(a).

  17. 17

    Artificial Intelligence Safety Measures Act § 18(a).

  18. 18

    Artificial Intelligence Safety Measures Act § 10(d).

  19. 19

    Artificial Intelligence Safety Measures Act § 17(a).

  20. 20

    Artificial Intelligence Safety Measures Act § 5.

  21. 21

    Artificial Intelligence Safety Measures Act § 5.

  22. 22

    Artificial Intelligence Safety Measures Act § 5.

  23. 23

    Artificial Intelligence Safety Measures Act § 5.

  24. 24

    Artificial Intelligence Safety Measures Act § 5.

  25. 25

    Artificial Intelligence Safety Measures Act § 25(c).

  26. 26

    Artificial Intelligence Safety Measures Act § 5.

  27. 27

    Artificial Intelligence Safety Measures Act § 5.

  28. 28

    Artificial Intelligence Safety Measures Act § 18(a).

  29. 29

    Artificial Intelligence Safety Measures Act § 18(b).

  30. 30

    Artificial Intelligence Safety Measures Act § 18(c).

  31. 31

    Artificial Intelligence Safety Measures Act § 5.

  32. 32

    Artificial Intelligence Safety Measures Act § 10(a).

  33. 33

    Artificial Intelligence Safety Measures Act § 10(a).

  34. 34

    Artificial Intelligence Safety Measures Act § 18(b).

  35. 35

    Artificial Intelligence Safety Measures Act § 10(e).

  36. 36

    Artificial Intelligence Safety Measures Act § 10(f)(1).

  37. 37

    Artificial Intelligence Safety Measures Act § 10(f)(2).

  38. 38

    Artificial Intelligence Safety Measures Act § 10(c)(1).

  39. 39

    Artificial Intelligence Safety Measures Act § 10(c)(2).

  40. 40

    Artificial Intelligence Safety Measures Act § 10(d).

  41. 41

    Artificial Intelligence Safety Measures Act § 10(d).

  42. 42

    Artificial Intelligence Safety Measures Act § 10(d).

  43. 43

    Artificial Intelligence Safety Measures Act § 10(g).

  44. 44

    Artificial Intelligence Safety Measures Act § 15(c).

  45. 45

    Artificial Intelligence Safety Measures Act § 15(c).

  46. 46

    Artificial Intelligence Safety Measures Act § 15(c).

  47. 47

    Artificial Intelligence Safety Measures Act § 15(f).

  48. 48

    Artificial Intelligence Safety Measures Act § 5.

  49. 49

    Artificial Intelligence Safety Measures Act § 20(a).

  50. 50

    Artificial Intelligence Safety Measures Act § 20(e)(1).

  51. 51

    Artificial Intelligence Safety Measures Act § 20(d).

  52. 52

    Artificial Intelligence Safety Measures Act § 25(a).

  53. 53

    Artificial Intelligence Safety Measures Act § 18(e).

  54. 54

    Artificial Intelligence Safety Measures Act § 25(b).

  55. 55

    Artificial Intelligence Safety Measures Act § 25(d).

  56. 56

    See Pritzker signs landmark AI regulation bill that aims to mitigate risks, Capitol News Illinois, July 6, 2026.

Contact