Illinois has imposed new constraints on how state and local governments can use artificial intelligence. For public agencies, vendors and IT teams responsible for procurement or operations, the changes create immediate compliance tasks and procurement considerations.
- Illinois restricted state and local government uses of AI, focusing on real-time biometric ID, public face recognition and continuous tracking.
- Agencies should inventory systems, pause certain procurements, and update contract terms to require vendor disclosures and audit rights.
- Vendors should prepare documentation about capabilities, training data and error characteristics and be ready to modify products or limit sales to public entities.
What the rules do — the essentials
The new guidance and restrictions limit specific uses of AI by state and local government entities. Key prohibitions target applications that perform real-time biometric identification or use face recognition to identify people in public settings. They also constrain AI deployment that tracks individuals across locations or makes high-stakes decisions about people without sufficient human oversight.
Because sources differ in wording and emphasis, agencies should treat this summary as a practical starting point and consult the official text—or legal counsel—before changing policy.
Who is affected
- State and local government bodies in Illinois that procure, develop or operate AI systems.
- Vendors and contractors supplying AI products or services to those agencies.
- IT, procurement, legal and records teams that handle contracts, data governance and public-facing systems.
Practical steps for agency leaders (priority checklist)
- Inventory AI uses now. Create a short register of all systems that use machine learning, computer vision, automated decision tools, data-matching or predictive analytics, including embedded third-party features in larger platforms.
- Flag biometric or face-recognition capabilities. Any tool that performs real-time identification or matches faces against a database should be classified as high-risk and paused pending review.
- Assess high-stakes decision systems. Identify systems that affect legal rights, benefits, employment, law enforcement or public safety outcomes and require human-in-the-loop controls and documented oversight.
- Review contracts and procurement language. Insert clauses requiring vendors to disclose AI capabilities, limitations, error rates, training data sources and update notifications where allowed by law. Require vendors to support audits and provide materials needed for oversight.
- Set a temporary procurement hold for prohibited uses. If your agency was planning to buy real-time biometric ID or continuous tracking tools, delay those purchases until compliance requirements are clear.
- Adopt simple transparency practices. Publish short explanations of when and why AI systems are used, what data they access, and a contact for questions or redress.
Vendor and supplier guidance
Vendors should expect stricter disclosure demands and should prepare documentation now. Useful preparations include:
- Detailed capability lists that clearly state whether a system performs real-time biometric identification, face matching in public spaces or continuous tracking.
- Technical documentation on model training data sources, approximate error rates and typical failure modes—presented in non-technical summary form for procurement teams.
- Contract templates that allow independent audits, limit re‑use of government data, and define responsibilities for compliance and incident response.
Why these changes matter — practical implications
For agencies, the rules narrow the set of permissible AI tools and increase the operational burden around procurement, documentation and oversight. For vendors, they create market friction: some products may no longer be sellable to Illinois public bodies without modification or stronger compliance guarantees.
From a citizen perspective, the restrictions aim to reduce privacy risks and the harms associated with automated identification in public places. But the rules also create trade-offs: agencies that relied on automated tools for efficiency may need to restore manual processes or invest in compliant alternatives.
Timeline and next steps
The available reporting indicates the rules are now in force for Illinois state and local government use. Agencies that have ongoing procurements or live systems using the restricted technologies should prioritize the inventory and contract review steps above.
If an agency needs to continue using an AI system that could be covered by the restrictions (for example, for public-safety operations), it should document the legal basis, scope limits, human oversight procedures and public notice measures supporting that exception.
Common questions agencies will ask
Does this ban all facial recognition?
No. The emphasis is on real-time identification in public spaces and systems that match faces against databases for identification without safeguards. Face-analysis functions used in controlled, consented settings or for limited administrative tasks may require a different compliance pathway—review the official guidance for specifics.
What about AI used for administrative automation (service routing, scheduling, chatbots)?
Routine automation tools that do not identify or track people and do not make high-stakes determinations are less likely to be affected. Agencies should, however, still document functionality and impacts and ensure human oversight where outcomes materially affect people.
Will vendors need to rebuild products?
Some vendors may only need to add disclosure and audit support; others whose products feature continuous public tracking or real-time biometric ID will need product changes or to accept limits on selling those features to Illinois public entities.
Unresolved issues and watchpoints
- Definitions. Practical questions remain about how narrowly regulators will define “real-time identification,” “tracking” and “high-stakes decisions.” Agencies should seek clarification from official sources or legal counsel.
- Federal interplay. How these state measures interact with federal guidance or procurement rules is not fully resolved; agencies receiving federal funds or operating under federal oversight should check for conflicts.
- Auditability. Vendors and agencies will need to test how audit requests and independent evaluations work in practice, especially where models rely on proprietary training data or external APIs.
Bottom line — immediate checklist
- Make an inventory of AI systems today.
- Pause procurements involving real-time biometric ID or continuous public tracking.
- Update procurement and contract language to require disclosures and audit access.
- Publish a short transparency notice for any AI system in public use.
- Consult counsel where systems make high-stakes decisions or law-enforcement uses are involved.
These steps are intended as practical, conservative measures to align operations with Illinois’s new limitations on government AI use. Agencies and vendors should consult the official text and legal advisors for binding guidance.
