Artificial Intelligence (AI) Transparency & Usage Policy

Purpose:

Mentavi Health uses artificial intelligence (AI) to improve your experience navigating our services — from answering questions about scheduling and billing to helping you understand our assessment process. This policy explains how AI is used in our patient- and user-facing systems, what it can and cannot do, and how we protect your privacy. We believe you deserve to know when you’re interacting with AI, what role it plays, and what safeguards are in place.

2. Scope

This policy applies to all AI-powered features you may encounter when using Mentavi Health’s services, including but not limited to:

• Our AI-powered support chatbot (available via our website and patient portal)
• AI-powered digital assistants, including visual or voice-enabled digital human interfaces
• Automated responses within our support and scheduling workflows
• Any future AI-assisted features introduced within Mentavi Health products This policy does not cover the clinical assessment instruments, diagnostic tools, or scoring methodologies used within our Diagnostic Evaluations or Mental Wellness Snapshots. Those are governed by separate clinical protocols and are always reviewed or supervised by licensed professionals. While AI assists with operational and informational tasks, human oversight is embedded at every level of these systems. Licensed professionals and compliance staff supervise AI interactions to ensure safety, accuracy, and adherence to Mentavi’s clinical standards — a framework described fully in Section 5.

3. AI Disclosure — You Are Interacting with AI, Not a Human

All AI-powered interactions at Mentavi Health — including text-based chatbots, voice-enabled assistants, and digital human interfaces — are artificial intelligence systems. They are not human and are not licensed mental health therapists, regardless of their visual or auditory presentation. Digital human avatars and voice-enabled AI used by Mentavi Health are computer-generated representations powered by language models. Their human-like appearance or voice does not indicate the presence of a real person. No human is speaking to you through these interfaces in real time. This disclosure is provided:

• Before you access any AI-powered feature. You will see a clear, prominent notice that the system is AI-powered before you can begin interacting with it. For digital human interfaces, this notice will appear before the avatar is presented.
• At the beginning of any interaction if you have not used the feature within the previous seven (7) days.
• Throughout your interaction. All AI interfaces maintain a persistent visual indicator that you are communicating with an AI system — not a human. For digital human avatars, this includes a permanent on-screen label identifying the interface as AI-generated.
• Verbally at the start of any spoken interaction. If an AI interface communicates with you using voice or audio, it will verbally identify itself as an AI system at the beginning of the conversation.
• Any time you ask. If you ask any Mentavi AI system whether you are interacting with AI, it will confirm that it is an AI system and not a human. Because Mentavi Health operates in a regulated healthcare environment and our AI interactions may involve health-related inquiries, we treat all AI interactions as high-risk and provide prominent disclosure proactively — regardless of whether you ask. These disclosure practices are designed to comply with Utah’s requirements for generative AI systems used in connection with regulated healthcare services, including proactive disclosure at the outset of AI-mediated interactions and truthful responses whenever a user asks whether they are interacting with AI.

4. How Mentavi Uses AI in Patient- and User-Facing Interactions

4.1 What Our AI Systems Do

Mentavi’s AI systems — whether presented as text-based chatbots or digital human interfaces — are designed to help you with non-clinical, operational tasks, including:

• Answering questions about Mentavi’s services (e.g., what’s included in a Diagnostic Evaluation vs. a Mental Wellness Snapshot)
• Providing information about pricing, scheduling, and availability
• Guiding you through account and portal navigation
• Explaining how our assessment and treatment processes work at a general level
• Directing you to the appropriate team member or resource when needed For existing patients, our AI support tools answer account, billing, and portal‑navigation questions and direct you to human staff. They do not interpret your test results, adjust your treatment, or provide clinical follow‑up. The underlying technology is the same regardless of the interface. A digital human avatar provides the same type of information and operates under the same restrictions as our text-based chatbot. The visual presentation does not change the system’s capabilities, limitations, or clinical boundaries.

4.1.1 Examples of AI Systems Covered by This Policy

• Gloria (Digital Human Assistant): A visual, voice-enabled digital human avatar deployed on Mentavi’s public websites to answer general questions for potential patients about Mentavi’s services, pricing, scheduling, and how to get started. Gloria does not provide clinical advice, conduct mental-health or ADHD evaluation, or access patient clinical records.
• Mentavi AI Support Application: A text-based AI assistant deployed on our public websites, patient portal, and support center to answer frequently asked questions for potential and existing patients, assist with account and portal navigation, and route users to human support representatives during Mentavi business hours. The Support Application does not provide clinical advice or therapy and escalates clinical questions to human staff.

4.2 What Our AI Systems Do NOT Do

All Mentavi AI systems are explicitly prohibited from:

• Providing clinical advice, diagnoses, or treatment recommendations. Our AI will not interpret symptoms, suggest diagnoses, recommend medications, or advise on treatment plans.
• Providing mental health therapy or counseling. Our AI is not designed to engage in therapeutic conversations, simulate the role of a licensed mental health therapist, or establish a therapeutic relationship — regardless of its visual or auditory presentation. Even when presented as a digital human avatar with a natural voice and expressions, our AI systems are not a therapist, are not a substitute for therapy, and are not designed to establish a therapeutic relationship.
• Conducting crisis assessments or suicide risk evaluations. If you indicate you are in crisis or experiencing a mental health emergency, the system will immediately direct you to appropriate crisis resources (see Section 8).
• Accessing, displaying, or modifying your Protected Health Information (PHI). Our AI does not have access to your clinical records, assessment results, or treatment history.
• Making decisions about your care. All clinical decisions at Mentavi are made by licensed psychologists and healthcare providers.
• Simulating empathy, emotional bonding, or a personal relationship. While our AI systems are designed to be respectful and responsive, they do not have emotions, do not form relationships, and are not a substitute for human connection or therapeutic rapport. These restrictions are enforced through technical safeguards built into the AI system (including content filtering and topic-denial guardrails), not just policy language.

4.3 Intended Purpose, Capabilities, and Limitations

• Intended purpose: Mentavi’s AI systems exist to provide operational and informational assistance to patients and prospective patients navigating Mentavi Health’s services. They are customer support tools, not clinical tools. This is true regardless of whether the interface is text-based, voice-enabled, or presented as a digital human avatar.
• Capabilities: Our AI can answer frequently asked questions, explain Mentavi’s service offerings, assist with scheduling and billing inquiries, and route you to the appropriate human team member.
• Limitations: Our AI cannot provide personalized medical or mental health advice, cannot access your individual health records, may occasionally misunderstand questions, and may provide information that is incomplete or not current. All AI-generated responses should be verified where accuracy is critical. The human-like appearance or voice of a digital assistant does not indicate greater intelligence, clinical expertise, or access to your personal information.

5. Human Oversight

AI at Mentavi operates under human supervision at all times:

• Licensed professionals review all clinical outputs. Our Diagnostic Evaluations are reviewed by licensed psychologists before results are delivered to you. AI does not make or influence clinical determinations.
• Licensed mental health professionals are involved in AI development and review. Our AI systems — including the content, tone, and behavioral boundaries of digital human interfaces — are developed and monitored with input from licensed clinicians to ensure consistency with clinical best practices.
• Support escalation is always available. If any AI system cannot adequately address your question, or if you prefer to speak with a human, you can request escalation at any time. Our AI is designed to recognize when escalation is appropriate and will proactively contact options, such as phone email, or live chat when available.
• Ongoing monitoring and review. Mentavi’s technical and compliance teams monitor AI interactions to ensure accuracy, appropriateness, and adherence to this policy. This includes regular, objective reviews of safety, accuracy, and efficacy.
• User safety prioritized over engagement. Mentavi prioritizes user mental health and safety over engagement metrics or profit. Our AI systems are designed to err on the side of caution and escalate to human support when in doubt. This applies equally to all AI interfaces, including those with more engaging visual or voice-based presentations.

6. Privacy & Data Protection

6.1 HIPAA-Equivalent Standards

Mentavi Health voluntarily holds itself to HIPAA-equivalent standards in the handling of all health-related information. All AI systems that interact with user data are designed and operated with these safeguards, including:

• Business Associate Agreements (BAAs) with all AI infrastructure providers
• Encryption of data in transit and at rest
• Access controls that limit AI system access to only the minimum necessary information
• Audit logging of AI interactions
• Compliance with the security and privacy provisions of 45 C.F.R. Part 160 and 45 C.F.R. Part 164, Subparts A, C, and E

6.2 Data Collection During AI Interactions

When you interact with any Mentavi AI system, the following may be collected:

• The content of your conversation (your questions/input and the AI’s responses), including text and, where applicable, voice input
• Session metadata (timestamps, device type, page of origin)
• Any information you voluntarily provide during the conversation (e.g., your name or email for follow-up) Voice and audio data from interactions with voice-enabled AI systems are processed for the purpose of understanding and responding to your queries. Voice data is not used to create voice profiles, identify you biometrically, or for any purpose other than processing your current interaction, unless you provide explicit consent.

6.3 What We Do NOT Collect Through AI Systems

• Clinical records, assessment results, or treatment details
• Payment card numbers or full financial account information
• Biometric identifiers (voiceprints, facial recognition) — our digital human interfaces present to you; they do not observe or analyze you
• Information beyond what you voluntarily provide in the conversation

6.4 No Sale or Sharing of Your Data

Mentavi Health does not sell individually identifiable health information collected through AI interactions to any third party. Mentavi Health does not share user input from AI interactions with any third party, except:

• When shared with a healthcare provider at your request and with your consent
• When provided to a health plan at your request
• When shared with a contracted service provider necessary for the AI system’s functionality, subject to appropriate privacy and security agreements equivalent to HIPAA business associate requirements

6.5 No Targeted Advertising Based on Your Input

Mentavi Health does not use your AI interactions to:

• Determine whether to display advertisements to you
• Select or customize advertisements based on your input or health information
• Target you with specific products or services based on what you share with our AI systems If our AI ever recommends that you seek counseling, therapy, or other assistance from a licensed professional, this is a safety and wellness referral — not an advertisement.

6.6 Data Retention & Use

• AI interaction logs are retained in accordance with Mentavi’s data retention policies and applicable legal requirements.
• Interaction data may be reviewed for quality assurance, safety monitoring, and service improvement purposes.
• Interaction data is never sold to third parties.
• De-identified, aggregated interaction data may be used to improve our AI systems and overall service quality.

6.7 Voluntary Disclosure Warning

Important: Please do not share sensitive clinical details, Social Security numbers, or payment card information with any Mentavi AI system. The conversational nature of these systems — particularly voice or avatar-based interfaces — may feel personal, but they are not private clinical sessions. If you inadvertently share sensitive information, contact us at [email protected] and we will take steps to remove it from our logs.

7. Reporting Harmful or Concerning Interactions

Mentavi Health is committed to identifying and mitigating potentially harmful AI interactions. If you experience an interaction with any Mentavi AI system that you believe is inaccurate, inappropriate, harmful, or concerning in any way:

• In-chat: Use the feedback mechanism within the interface to flag the specific response.
• By email: Contact [email protected] with a description of the interaction.
• By phone: Call [phone number] during business hours. All reports are reviewed by Mentavi’s compliance and technical teams. We use reported incidents to improve safety, accuracy, and the overall quality of our AI systems. Mentavi maintains documented protocols for assessing and responding to risk of harm to users, including protocols for real-time response to acute risk of physical harm.

8. Crisis & Safety Protocols

No Mentavi AI system is a crisis service, and no AI interaction is a substitute for emergency medical or mental health care — regardless of how lifelike the interface may appear. If you are experiencing a mental health emergency, please contact:

• 988 Suicide & Crisis Lifeline: Call or text 988 (available 24/7)
• Crisis Text Line: Text HELLO to 741741
• Emergency Services: Call 911 If any Mentavi AI system detects language that may indicate a crisis, it will:

1. Pause the current conversation flow
2. Provide crisis resource information immediately
3. Offer to connect you with a Mentavi team member This response is automated and intentionally errs on the side of caution. We would rather provide resources you don’t need than miss an opportunity to help.

9. Accuracy & Limitations

9.1 AI-Generated Content May Contain Errors

AI systems, including ours, can produce responses that are inaccurate, incomplete, or outdated. While we work to ensure our AI provides reliable information:

• Responses about Mentavi’s services, pricing, and processes are based on curated, approved content and are regularly reviewed for accuracy.
• Our AI may occasionally misunderstand your question or provide a response that doesn’t fully address your needs.
• AI-generated responses should not be treated as medical, legal, or financial advice.
• The confidence or fluency of an AI response — whether delivered as text, voice, or through a digital human avatar — is not an indicator of accuracy.

9.2 Accountability for AI Outputs

Mentavi Health is responsible for the outputs of its AI systems. The use of AI is not a defense for any statement or act that would otherwise constitute a consumer protection violation. Mentavi treats AI- generated communications with the same level of accountability as statements made by its employees.

10. Non-Discrimination

Mentavi Health implements measures to prevent discriminatory treatment of users by our AI systems. Our AI systems are monitored for bias and fairness, and we are committed to equitable treatment of all users regardless of race, ethnicity, gender, sexual orientation, age, disability, or any other protected characteristic.

11. Your Rights

As a Mentavi Health patient or user, you have the right to:

• Know when you’re interacting with AI. All AI interactions are clearly identified, with disclosures provided before access, upon return after seven days, verbally at the start of spoken interactions, and whenever you ask.
• Request human assistance at any point during an AI interaction.
• Opt out of AI interactions and receive support through non-AI channels (email, phone).
• Request information about how your data from AI interactions is used.
• Request deletion of your AI interaction data, subject to legal retention requirements.
• Report harmful interactions through the mechanisms described in Section 7.
• Provide feedback on AI interactions to help us improve.

12. Changes to This Policy

Mentavi Health may update this policy as our AI capabilities evolve or as regulatory requirements change. Material changes will be communicated via our website and, where appropriate, direct notification. We encourage you to review this policy periodically.

13. Contact Information

For questions about this policy or Mentavi’s use of AI:

• Privacy & Compliance: [email protected]
• General Support: [email protected]
• Mailing Address: [Mentavi Health address]

14. Related Policies

• Mentavi Health Privacy Policy (Notice of Privacy Practices)
• Terms of Service
• Website Cookie & Tracking Policy
• Internal AI Chatbot Usage Policy
• AI Governance and Ethics Committee Charter

Mentavi Health is committed to the responsible use of artificial intelligence. Our AI systems are designed to support — never replace — the licensed professionals who deliver your care.