Illinois Weighs Liability Limits for Powerful AI Models

New legislation aims to balance innovation and safety as companies warn of AI risks.

Apr. 10, 2026 at 2:07pm

Got story updates? Submit your updates here. ›

A highly detailed 3D illustration of a glowing, futuristic AI neural network, with pulsing neon blue and purple lights flowing through intricate circuits and data pathways, suspended in a dark, atmospheric environment to convey the power and complexity of advanced AI models.As Illinois lawmakers consider new liability limits for powerful AI systems, the cybersecurity risks and potential for catastrophic harm posed by these advanced models come into focus.Springfield Today

The Illinois state legislature is considering a bill that would create a framework to shield developers of advanced artificial intelligence models from liability for catastrophic harms, as long as they meet strict transparency and safety requirements. The proposed Artificial Intelligence Safety Act would apply only to the most powerful and expensive AI systems, not consumer products or business software.

Why it matters

As AI capabilities continue to advance, policymakers are grappling with how to encourage innovation while also mitigating the potential for these systems to cause large-scale damage. This legislation represents an attempt to find a middle ground by providing legal protections for companies that proactively address AI safety concerns.

The details

Senate Bill 3444 would require developers of frontier AI models, defined as those exceeding $100 million in compute expenses, to publish safety protocols and risk assessment reports before release. In exchange, they would generally be shielded from liability for catastrophic outcomes like mass casualties or billion-dollar property damage, unless they intentionally or recklessly caused harm. The bill allows companies to satisfy the requirements by complying with EU standards or entering evaluation agreements with federal agencies.

  • The Illinois state legislature is currently considering Senate Bill 3444.
  • If passed, the law would automatically stop applying if Congress enacts overlapping federal AI regulations.

The players

Sen. Bill Cunningham

The Democratic state senator who introduced Senate Bill 3444, the Artificial Intelligence Safety Act.

Anthropic

An AI research company that recently announced it would delay the release of its latest model, Claude Mythos Preview, over concerns about its cybersecurity capabilities.

Got photos? Submit your photos here. ›

What’s next

If passed, the Artificial Intelligence Safety Act would establish a first-of-its-kind framework in Illinois to address the risks of powerful AI models. The legislation would need to be monitored for any updates or changes as it moves through the state legislature.

The takeaway

This proposed legislation in Illinois represents an attempt to balance the need for AI innovation with the imperative to ensure these advanced systems do not pose unacceptable risks to public safety. By providing liability protections for companies that meet strict transparency and safety requirements, the state is seeking a middle ground that could serve as a model for other jurisdictions grappling with the challenges of regulating transformative AI technologies.