0
Skip to content

Illinois Passes Historic AI Safety Act

gavel-ai-tech-law-court-dataIllinois is stepping into the AI legislative arena with SB 315. While drawing inspiration from existing AI transparency rules in California and New York, Illinois’ SB 315 goes a step further by mandating a strict, legally binding third-party audit system.

6/5/2026 1:00 p.m.

The Illinois Legislature recently passed SB 315, a landmark bill targeting the creators of advanced frontier AI models.

Passing the House with a unanimous 110-0 vote and the Senate 52-5, the bill is now headed to Governor JB Pritzker, who has indicated he plans to sign it, according to a report from NBC News.

While the legislation directly targets major AI developers (like OpenAI and Anthropic) rather than the end-users of AI, the ripple effects could reshape the compliance, vendor risk management, and operational landscapes for the financial services sector.

SB 315 builds on existing AI transparency frameworks in states like California and New York by introducing a strict, legally binding audit mechanism.

The legislation requires frontier AI developers to:

  • Submit to Third-Party Audits: In a U.S. first, developers must undergo annual, independent third-party safety audits to evaluate severe or catastrophic risks.
  • Publish Public Safety Plans: Companies must create, publish, and annually update frameworks detailing how they mitigate catastrophic risks.
  • Protect Whistleblowers: The law establishes formal reporting processes and robust protections for AI company employees who flag safety issues.

The law is slated to take effect on Jan. 1, 2027, with violations carrying steep civil penalties.

Financial Services Takeaways

Even if your firm isn’t building frontier large language models from scratch, this legislation impacts the broader financial landscape in three distinct ways:

  • A New Baseline for Vendor Due Diligence: As financial institutions increasingly integrate AI into operations and customer service, vendor risk assessment is becoming highly complex. Once SB 315 takes effect, compliance and risk officers will have access to a standardized repository of independent safety audits and public risk frameworks from major AI providers.
  • The Fragmented Regulatory Burden: Because federal AI regulation remains stalled in Congress, states are aggressively stepping into the vacuum. The White House has previously expressed concern that a patchwork of state-by-state laws could complicate corporate compliance. For national financial firms, keeping track of varying AI governance baselines across Illinois, California, New York, and others will require a more agile and sophisticated compliance architecture.
  • Market Stabilization and Tech Buy-In: The bill has received public support from major AI developers like OpenAI and Anthropic, who view it as a way to institutionalize the safety frameworks they already practice voluntarily. For the financial sector, mandatory safety baselines could actually accelerate institutional adoption by lowering the perceived risk of deploying advanced AI tools.
New AI Models and ARM Industry Impact

Meanwhile, ACA International recently reported that Anthropic has announced a new AI model, Mythos, which researchers say has the capacity for autonomous cyber operations far exceeding those of any previous tool. During internal testing, the model successfully bypassed security protocols and identified thousands of high-severity vulnerabilities across all major operating systems and web browsers.

According to Anthropic, the model has reached a level of coding proficiency that allows it to outperform all but the most skilled humans in finding and exploiting software flaws, even self-correcting its own errors during the process.

The discovery of these capabilities has triggered immediate concern at the highest levels of the U.S. government.

State regulators are also taking note.

For example, the New York State Department of Financial Services (DFS) issued new industry guidance outlining operational steps financial institutions should consider to protect data and infrastructure during heightened cybersecurity threats, ACA recently reported.

The DFS believes that the release of a highly advanced, large-scale AI tool, known as a “frontier model,” should alert companies.

For the accounts receivable management (ARM) sector, which handles massive volumes of personally identifiable information, the existence of an AI model capable of sophisticated hacking represents a significant shift in the threat landscape.

Regarding state and federal AI regulations for financial services and the private sector, ACA continues to advocate for a tailored approach to governing the technology’s uses in the ARM industry.

ACA has urged Congress and federal regulators to recognize AI as essential for consumers’ flexibility in managing account information and payments, emphasizing that national AI priorities should reflect these benefits for consumers and businesses.

Related Content from ACA International:

White House Releases National AI Framework to Streamline Regulation

ACA Huddle Recording Available — Behind the Prompt: AI, Privilege & Discovery Risks

Remember to subscribe to ACA Daily and Member Alerts under your My ACA Assistant profile when logged in to acainternational.org.

If you have executive leadership updates or other member news to share with ACA, contact our communications department at comm@acainternational.org. View our publications page for more information and our news submission guidelines here.

One moment please...

Share Profile

This site utilizes cookies as included in our Privacy Policy and User Agreement.