AI Disclosure Requirements Under Colorado SB 24-205

Under SB 24-205, deployers must provide consumers with specific information when AI is used in consequential decisions:

• AI involvement — That an AI system is being used to make or substantially influence the decision
• Decision type — What type of consequential decision is being made (employment, credit, housing, etc.)
• Appeal rights — How the consumer can contest the decision or request human review
• Policy access — Where to find the deployer's AI risk management policy
• Contact information — How to reach the organization with questions or concerns

The disclosure must be in "plain language" — no legal jargon, no technical terminology that a typical consumer wouldn't understand.

Disclosure timing is critical and often misunderstood. The law requires notice before or at the time the AI system interacts with the consumer. Not after. Not buried in a follow-up email.

Practical examples:

• AI chatbot on your website — Disclosure before the conversation begins. A banner or initial message stating "This chat uses AI to help direct your inquiry."
• AI-powered hiring tool — Disclosure in the job application process before the AI evaluates the candidate. Include it in the application portal, not the offer letter.
• AI credit scoring — Disclosure at the point of application, explaining that AI will be used in the lending decision.
• AI legal research tools affecting case strategy — Disclosure to clients when AI materially influences decisions about their case.

The statute requires disclosures to be "clear and conspicuous." Based on FTC guidance and emerging AG interpretation, this means:

• Prominent placement — Not hidden in page 37 of your terms of service. The disclosure must be where consumers will actually see it.
• Plain language — Write at an 8th-grade reading level. "We use artificial intelligence to evaluate your application" beats "algorithmic decision-making systems are employed in the assessment pipeline."
• Accessible format — Mobile-friendly, ADA-compliant, available in the consumer's language when practicable.
• Standalone notice — Don't bundle it with 50 other disclosures. AI use disclosure should be distinct and separately acknowledgeable.

Three proven disclosure implementation patterns:

Pattern 1: Inline Banner
A persistent, visible banner at the interaction point. Example: "AI Notice: This service uses artificial intelligence to [specific action]. You have the right to request human review. Learn more."

Pattern 2: Pre-Interaction Gate
A disclosure screen before the AI interaction begins. The consumer must acknowledge the disclosure before proceeding. Best for high-stakes decisions (hiring, lending).

Pattern 3: Contextual Notice
AI disclosure integrated into the workflow at the moment it becomes relevant. When AI generates a recommendation, the output includes a notice: "This recommendation was generated by AI. Request human review."

CO-AIMS provides pre-built disclosure templates for each pattern, automatically customized to your industry and AI systems. Disclosures are logged with timestamps for your compliance records.