AI Policy

SEER 44 - Use of Artificial Intelligence

DIVISION:	Agency Wide
DEPARTMENT:	All
EFFECTIVE DATE:	05/29/2026
REPLACES POLICY DATED:	07/01/2025

PURPOSE:

The emergence of artificial intelligence (AI) has presented new opportunities that can be a transformational enhancement to the work that we do at Orchards Children’s Services. While there are opportunities for positive AI usage, there can also be ethical considerations that staff must consider. Some of the followings are areas for consideration:

Accountability AI can provide recommendations, but the responsibility for decision-making and accountability is still the responsibility of Orchards Children’s Services staff. There must be staff oversight for AI-generated plans or interventions. The use of an assessment tool cannot be replaced using AI.

Client Welfare It is always the responsibility of Orchards Children’s Services staff members to ensure we have our clients’ best interests in the delivery of services. If our client’s well-being, confidentiality, or outcomes could be compromised using AI, alternative approaches should be used.

Competence (AI) Prior to using AI, staff must possess a foundational knowledge of how the technology works, its capabilities and limits, and the confidentiality of client data. Any time a tool, assessment, or intervention is used with a client, the staff member has an ethical obligation to know how to appropriately use the tool.

Competence (Service Delivery) Orchards Children’s Services staff may not use AI as a replacement for professional judgment. Staff competence in assessing, decision-making, and intervention planning is still essential. Consideration must also be given to ensure that multicultural issues are considered, and there is no bias in decision making. It must be acknowledged that bias has been found and documented in AI responses, and DEI knowledge must be applied to combat that bias.
Confidentiality We have an obligation to ensure that client information is kept confidential by avoiding the use of any identifiable information when AI is being used. Staff should also be aware that most AI sites do not have a confidentiality clause in place. Even sites who claim to have confidentiality are not to be considered confidential by Orchards Children’s Services staff as this claim has been questioned. AI sites remember and then use with others information you submit into the sites, so organization intellectual property should not be entered into an AI site.

Definitions:

Artificial Intelligence (AI): Technology that simulates human intelligence processes, including machine learning, natural language processing (NLP), and generative AI (e.g., text, image, audio generation).
AI Tool: Any software or platform that incorporates AI capabilities for content generation, prediction, automation, or decision support. Orchards Children’s Services uses Microsoft Copilot as the agency’s official AI tool.
PII (Personally Identifiable Information): Is any information connected to a specific individual that can be used to uncover or steal that individual's identity, such as their social security number, full name, email address or phone number.
PHI (Protected Health Information): Is any health-related data that can be linked to a specific individual. Under HIPAA (Health Insurance Portability and Accountability Act), data becomes PHI if it is created or maintained by a healthcare provider, health plan, or healthcare clearinghouse, and it contains both health information and any standard identifiers.

POLICY:

This policy applies to all Board members, staff, volunteers, interns, and any other individuals who have access to and use AI systems on behalf of Orchards Children’s Services. This policy also covers the use of AI tools and applications for work-related tasks, both on-site and remote working. Orchard’s uses Microsoft Copilot as the agency’s authorized AI tool. The use of all other AI tools and platforms including but not limited to ChatGPT, DeepSeek, Gemini, etc., is strictly prohibited.

Legal and Regulatory Compliance:

All use of AI must comply with applicable federal, state, and local laws, including but not limited to HIPAA, and relevant data protection regulations. Use of AI that violates any regulatory requirements or client rights will be subject to disciplinary action.

PROCEDURE:

Acceptable Use:

The acceptable use of AI is about advancing technology to increase efficiency and quality of services while also using our ethical obligations to protect our clients. These are the principles that should define the acceptable use of AI at Orchards Children’s Services:

Strengthening Service Delivery – AI can be used to improve the quality and effectiveness of services provided to clients. This can include administrative tasks, increasing efficiency, enhancing document management, automating repetitive tasks, making data driven decisions, and recommending process improvements.
Identifying Bias to Ensure Equity – Staff should be aware of common biases that could be present within AI, such as data bias, algorithmic bias, and minimizing disparities in outcomes with clients from diverse populations. AI systems are biased toward Standard English which could lead to bias with other cultures and or ethnic groups such as African American or other Indigenous groups. When reviewing reports, supervisors will be mindful of assessing and mitigating potential bias.
Protecting Privacy and Confidentiality – AI systems should only be used for low level data and not include any client identifying information. Any data breaches involving AI systems must be reported to the Supervisor, IT Department, and the Human Resources and Compliance Director.
Ensuring Accuracy and Appropriateness of AI-generated Outputs – It is important to verify the accuracy of AI generated content by cross-checking with reliable sources including the use of human judgment or other reliable sources. Review content for appropriateness by considering the context, audience, and impact of the content. At times AI systems can be misleading because the output is the most used data in a dataset. This means it could suppress information that is not as frequent and/or marginalized information.
Human Oversight – AI tools should be used with the end-user and client at the center of the decision-making process. AI is designed to enhance and not replace human interactions when delivering services.
Training and Awareness – It will be important that staff members familiarize themselves with the AI policy and stay up to date with trends and relevant best practices when it comes to AI usage. Continuous learning and professional development will help ensure responsible use of AI.

Unacceptable Use:

Using AI to make employment decisions (hiring, promotions, etc.), generating misinformation, or violating data privacy regulations. Use of AI as a stand-alone assessment tool.
Using AI for tasks that require sensitive information or involve legal compliance. Inputting client PHI, PII or other sensitive information during the use of AI.
Using AI that puts the agency’s or DHHS data, intellectual property, or confidential information at risk.

MONITORING PLAN:

Evaluation/Assessment of AI Tools:

When using AI systems, staff should consider the following questions:

Does using an AI system enable or hinder my ability to do my job effectively?
Is the content generated verifiable and accurate?
Will using AI systems negatively impact our clients and/or additional stakeholders?
Is the content that is generated from bias that could harm others?

Staff Training:

This policy is shared with newly onboarded staff on or before their first day of employment.

Staff are trained annually on the acceptable use of AI including any industry updates.

Access Control:

Do not share log in credentials or sensitive information with others or third parties. Staff should follow the agency’s Acceptable Use Policy CRP-25 as it applies to passwords and the security of data.