EPIC, CFA, Fairplay Submit Recommendations Ahead of Potential Rulemaking for Colorado Automated Decision

Updates
EPIC, CFA, Fairplay Submit Recommendations Ahead of Potential Rulemaking for Colorado Automated Decision-Making and Chatbot Laws
July 13, 2026
EPIC, the Consumer Federation of America, and Fairplay submitted comments on Monday in response to the Colorado Department of Law’s request for stakeholder input ahead of potential rulemaking for the state’s recently passed bills on chatbot safety and automated decision-making technology (ADMT) in consequential decisions.
ADMTs and chatbots both introduce new vectors of risk for Coloradans that often occur behind the shadow of the corporate “black box.” But as our comments explain, with strong, consumer-minded rulemaking, the Department can better protect Coloradans by closing some of the loopholes in the legislative text.
For the ADMT bill, EPIC, CFA, and Fairplay recommended that the Department use the rulemaking process to:
- broaden the scope of “consequential decision;”
- clarify what qualifies as an “adverse outcome;”
- offer specific data disclosure guidance;
- clarify the manner by which deployers must provide “clear and conspicuous” notice to consumers;
- require risk assessments prior to ADMT deployment; and
- clarify the limited scope of developers’ and deployers’ “right to cure” violations of law.
Regarding the Chatbot Safety Act, we encouraged the Department to:
- clarify and carefully tailor the scope of the bill’s exemptions to prevent companies from exempting themselves from the law;
- institute a process to approve new, privacy-protective age determination tools in the future;
- require meaningful safety reports to inform the public about how companies are moderating their chatbots and how people are using them; and
- increase accountability and transparency into how companies are responding to serious risks to protect consumers without putting consumers’ personal information at risk.
EPIC has been engaged in advocacy around Colorado’s law on automated decision-making technologies since its original passage in 2024. EPIC, CFA, and Fairplay have all been calling for strong chatbot regulations across the country and collaborated to develop the People-First Chatbot Bill, which was recently introduced in Congress.
Support Our Work
EPIC's work is funded by the support of individuals like you, who allow us to continue to protect privacy, open government, and democratic values in the information age.
Donate

Facts Only

* EPIC, CFA, and Fairplay submitted comments to the Colorado Department of Law.
* The comments were in response to a request for stakeholder input regarding potential rulemaking for state bills on chatbot safety and automated decision-making technology (ADMT).
* Recommendations for the ADMT bill included broadening the scope of "consequential decision."
* Recommendations also included clarifying what constitutes an "adverse outcome" and offering specific data disclosure guidance.
* Other recommendations covered clarifying notice requirements, risk assessments prior to deployment, and defining the scope of the right to cure violations.
* For the Chatbot Safety Act, recommendations included clarifying exemptions, establishing a process for approving age determination tools, requiring safety reports on chatbot moderation, and increasing accountability regarding risk response.
* EPIC has advocated for Colorado’s law on automated decision-making technologies since its passage in 2024.
* The organizations collaborated to develop the People-First Chatbot Bill.

Executive Summary

EPIC, the Consumer Federation of America, and Fairplay submitted comments to the Colorado Department of Law regarding potential rulemaking for state bills concerning chatbot safety and automated decision-making technology (ADMT). The organizations recommended that the Department use the rulemaking process to broaden the scope of "consequential decision," clarify what constitutes an "adverse outcome," provide specific data disclosure guidance, define how deployers must give consumers "clear and conspicuous" notice, require risk assessments before ADMT deployment, and clarify the limited scope of the right to cure for developers and deployers. For the Chatbot Safety Act, they urged clarification of exemptions, the establishment of a process for approving privacy-protective age determination tools, mandatory safety reports on chatbot moderation and usage, and increased accountability regarding risk mitigation without compromising personal information. The organizations have advocated for strong chatbot regulations nationally and collaborated on the People-First Chatbot Bill.

Full Take

The submission reflects an effort to shift a regulatory process from legislative text to concrete, enforceable consumer protection standards by focusing on the mechanisms of risk. The recommendations target the opacity inherent in "black box" systems by demanding specific definitions for outcomes and remedies, suggesting that without explicit guidance, current legislation leaves significant regulatory gaps. A pattern emerges where advocacy groups seek to operationalize abstract legal concepts—like "consequential decision"—into measurable requirements, effectively attempting to insert a rights-based framework into technology governance. The focus on risk assessments and mandatory reporting suggests an underlying assumption that the potential harm from ADMTs and chatbots necessitates proactive, pre-deployment oversight rather than reactive litigation. This effort reflects a broader societal pattern where decentralized technological power is being countered by organizing efforts demanding transparency and accountability in algorithmic decision-making. What remains to be explored is whether this level of specificity can successfully navigate the tension between innovation incentives and consumer protection mandates without stifling deployment.

Sentinel — Human

Confidence

This text appears to be a summary of stakeholder input for a regulatory process, exhibiting the characteristic structure of advocacy communications rather than pure news reporting.

Signals Detected

Moderate sentence length variance and appropriate use of transitional phrases.

Clear argumentative structure focused on recommendations, exhibiting a specific advocacy tone.

Direct reference to specific organizations (EPIC, CFA, Fairplay) and legislative processes suggests grounded sourcing.

No immediate signs of manufactured statistics or highly polished, impersonal prose; content reads like a press release summary.

Human Indicators

The text effectively frames a policy advocacy effort by citing specific organizational inputs and recommendations, which requires human engagement with legal/policy material.

The tone shifts between objective reporting of submitted items and persuasive framing (e.g., 'better protect Coloradans') which suggests intentional rhetorical choice.