Artificial intelligence tools present both opportunities and risks for community advocacy organizations. WHNC supports responsible AI use much like it is looking towards sustainable future housing growth for California via responsible development. The responsible use of AI is important as both 1) a cost savings measure and 2) a force multiplier on Community Funds and Volunteer Efforts. This policy establishes guidelines to ensure that AI is used responsibly, transparently, and in alignment with our commitment to serving the Woodland Hills community.

AI Use

AI tools are utilized for both internal operations and external community engagement. Internally, AI supports staff and volunteers in analyzing land-use and development project information, public safety data and incident reports, and environmental data including air quality and traffic patterns. Externally, AI may be used in public-facing communications and community engagement platforms. All AI applications serve to enhance, not replace, human judgment in advocacy decisions.

Themes

We see three important themes: disclosure, data and information protection, and content responsibility.

1. Disclosure: Be transparent in the disclosure of the use of AI so others are clearly aware

2. Data and Information Protection: We do not put anyone’s personal information into Gen AI tools. Gen AI tools do not have access to our mailing list.

   We are mindful of our user rights regarding data usage by the AI - this means, we opt out of our data being used to train AI models (LLM training) Specifically, we use a Nonprofit enterprise version of Gemini in a paid Google workspace - data stored there is private and never used for LLM model training). Our other tools require us to intentionally opt out of data sharing and model training even for paid accounts - and we do so. If you want to follow suite, OpenAI / Chat GPT allows this via their Settings > Data controls on individual accounts, JellyPod for audio creation allows has privacy settings to turn off data sharing & model training, and Anthropic (Claude) you must manually opt out in settings or your data could be retained for up to 5 years.

3. Content Responsibility: Since GenAI systems are fallible, in addition to #1 and #2 above - we have multiple people review all GenAI output before it is distributed to any audience content (text, graphic, audio, video) for correctness and clarity.

Interactions with Beneficiaries

When AI tools interact with community members, clear disclosure is provided regarding their use. AI-generated insights and analyses are made available to community members upon request to ensure accessibility and accountability. Public meetings and official communications include disclosures when AI has been used in developing advocacy positions or analyzing community data.

Data Collection Practices

Only publicly available data that does not require individual consent is used in AI systems. This includes land-use and development project information, public safety data and incident reports, and environmental data such as air quality and traffic patterns. No personal data requiring explicit consent is collected or processed through AI tools.

EMAIL NOTE

Ethical Risks and Concerns

Three primary ethical concerns guide AI use: potential bias in analyzing community data that could misrepresent certain neighborhoods or populations, transparency issues about how AI influences advocacy positions or recommendations, and accuracy of AI-generated information that could affect land-use decisions. Regular audits of AI outputs are conducted to identify potential biases, and multiple data sources are used to cross-check AI findings before they inform advocacy work. Human review is required for all AI-generated insights before they are incorporated into public positions or recommendations.

Accountability for AI Decisions

The Board of Directors maintains ultimate accountability for all decisions made with the assistance of AI tools. AI serves only as a support tool, with final decisions resting with organizational leadership. No automated decisions are made without human oversight and approval.

Community Feedback

A dedicated channel is provided for residents to share feedback or concerns about AI use in the organization's advocacy work. This ensures ongoing community input and allows for responsive adjustments to AI practices based on resident concerns and preferences.

Third-Party AI Tools

Third-party AI tools such as chatbots, data analytics platforms, or mapping software are used or planned for use. All third-party tools are evaluated for compliance with this policy before adoption, including assessment of their data practices, transparency, and potential for bias. Contracts with third-party vendors include provisions requiring adherence to the organization's ethical standards and data handling practices.

Staff Training

Staff and volunteers who use AI tools receive basic orientation on how to use specific AI tools effectively. Training ensures that users understand the capabilities and limitations of AI systems and can apply them appropriately in their work.