From the State Government edition of October 28, 2026
Colorado AG hearing on ADMT and Chatbot Safety Act rules (4 CCR 904-6): business groups seek delay and narrower definitions; advocates oppose further weakening
Colorado Department of Law (Attorney General), as reported by The Circle Desk (Legal Tech Circle) · State regulatory rulemaking; covered domains include essential government services · Colorado, United States
- Publisher
- Legal Tech Circle
- Original publication
- 28 October 2026 (Legal Tech Circle); hearing 26 October 2026
- Source retrieved
- 2026-10-29
- Event date
- 2026-10-26
What happened
The AG held a formal hearing on 26 October on 4 CCR 904-6, which implements SB 26-189 (ADMT Act) and HB 26-1263 (Chatbot Safety Act). Both take effect on 1 January 2027, and the AG must adopt rules before then. Business groups (U.S. Chamber, lenders, health care providers, Colorado Technology Association) asked for a one-year enforcement delay or good-faith safe harbor and narrower definitions of "material influence," "data sources" and "meaningful human review." EPIC, TechEquity and others urged the office to hold the line, objecting to the "commercially reasonable" and "technically feasible" qualifiers and to the October draft's removal of mandatory stays of decisions during review. The covered domains include essential government services. Deployers must give notice before use, explain adverse outcomes within 30 days, and offer correction and human review where commercially reasonable. No date for final rules. Under a court order in xAI's suit, enforcement waits until 14 days after a ruling on the preliminary injunction.
Why it matters
This resolves an open watch from this stream (hearing outcome): no final rules yet, and the testimony shows that "meaningful human review" is the contested control. That matters for any state agency providing essential government services in Colorado and for other states copying the model.
Evidence and measured results
Read the full Legal Tech Circle article. It is a secondary synthesis citing The Sum and Substance, Bloomberg Government, Law360, Finnegan, Jones Walker, the AG rulemaking page and the SOS docket; those originals were not opened. Whether and how Colorado state agencies are treated as deployers was not analysed in the source.
Limitations and uncertainty
Secondary aggregator source from a legal-tech publication. Testimony details come second-hand. Final rules pending. Litigation (xAI) may delay enforcement.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-10-28; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- Organizations covered by Colorado's ADMT Act, including essential government services, face a 1 January 2027 compliance date with rule definitions still contested.
- Stakeholders
- Agency general counsel
- Program directors for eligibility or licensing
- CIO
- Vendor product managers
- Compliance officer
- Discovery questions
- Which decisions does AI materially influence?
- Can you explain adverse outcomes within 30 days?
- How is meaningful human review defined in your process?
- How will the final rule text change your plans?
- Value hypothesis
- Early inventory and templates could reduce rush work once rules are final. No measured benefit; depends on final text.
- Potential engagement
- A 4-week ADMT gap assessment for one program, refreshed when the final rules are adopted.
- Unsupported claims
- Final rule content
- Enforcement date certainty
- That state agencies are definitively covered in every use
- Compliance cost estimates
Pre-sales engineering
Role takeaway
- Fit
- Fits decision-support systems in eligibility, licensing, employment or lending workflows subject to Colorado law.
- Architecture and integration
- Add notice, explanation, data-correction and human-review routing to case-management workflows with decision lineage logging.
- Prerequisites
- Inventory of consequential decisions
- Data lineage
- Configurable workflow engine
- Template management
- Constraints
- Rules not final
- Contested definitions
- Litigation-dependent enforcement timing
- Security
- Protect decision records and data-source lineage created for explanations and correction requests.
- Proof of value
- Generate a test adverse-outcome explanation from logged data sources
- Route a sample decision to human review and measure time
- Toggle thresholds for material influence
Delivery
Role takeaway
- Work
- Inventory decisions, draft templates, design review queues and monitor rule adoption and litigation.
- Dependencies
- Final 4 CCR 904-6 text
- Counsel interpretation
- System vendors
- Program staff
- Ownership
- Program owner with general counsel.
- Skills and adoption
- Train reviewers on meaningful human review and explanation writing.
- Governance checkpoints
- Final rule adoption
- xAI injunction ruling
- Pre-January readiness review
- Acceptance criteria
- Proposed: 100% of covered decisions inventoried
- Proposed: explanations within 30 days in a test sample
- Proposed: review queue staffed per policy
- Risks
- Rules change late
- Over-compliance cost
- Reviewer capacity shortfalls
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Systems that materially influence consequential decisions will need notice, adverse-action explanation and data-correction workflows; interpretation: state agency case-management systems may need these capabilities by 2027 if they are covered.
Governance
Who approves, reviews and stays accountable for outcomes?
The removal of mandatory decision stays during review and the commercially reasonable qualifiers weaken human-review guarantees. The final rule text will set the floor that other states may cite.
Security and privacy
What data, permissions and controls need testing?
Data-correction rights require reliable identification of the data sources used in decisions, which also creates records that need protection.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
Human-review duties need trained reviewers; the scope ("where commercially reasonable") will determine staffing needs.
Procurement
What should contracts, pricing and exit terms secure?
Vendors of decision-support tools sold to Colorado deployers, potentially including agencies, will need notice and explanation features and documentation of data sources.
Operating model
Which teams own the service once it runs?
Organizations need an owner for adverse-outcome explanations within 30 days and a human-review queue.
Publication history
- 2026-10-28State Government · Issue 514 resources
Stable resource ID: colorado-ag-admt-chatbot-rules-hearing-testimony-20261028